Brian R. King https://www.eg.bucknell.edu/~brk009/blog.html Associate Professor of Computer Science, Bucknell University — AI, machine learning, data science, and XR. quarto-1.10.19 Tue, 09 Jun 2026 04:00:00 GMT Workshop – Beyond Prompting: A Gentle Introduction to Agentic AI – May 21, 2026 https://www.eg.bucknell.edu/~brk009/posts/2026-06-09-workshop-beyond-prompting-a-gentle-introduction-to-agentic-ai-may-21-2026/ Agentic AI refers to artificial intelligence systems that can act autonomously — planning, making decisions, and executing multi-step tasks with minimal human intervention. Unlike traditional LLM-based tools that simply respond to a single prompt, agentic systems can complete tasks that you design. You can design an agent to browse the web, run and debug code, manage your files, review and send emails, and chain together complex workflows to accomplish your goals on your behalf.

For faculty and staff, understanding Agentic AI is of great interest for a variety of reasons. Faculty want to understand it to a) see what their students are doing with AI, and b) see if there are any opportunities to improve their own workflows in a safe, secure way. Of particular interest is the use of AI systems to improve research. (Your favorite chatbot can be a pretty decent “research assistant” if guided properly!) Likewise, administrative staff want to understand how their tasks might be streamlined. And, most importantly for us, we all want to know how our students’ learning is affected when they engage with these tools. (And if you think they aren’t using AI, you’d be wrong.)

Knowing what agentic tools can do and where their limits and risks lie helps educators make informed decisions about how we might integrate them responsibly into teaching, research, and daily workflows. It also helps us understand when not to use them. As these technologies become embedded in mainstream software and institutional processes, fluency with agentic AI is fast becoming a core professional skill.

To help our community address these urgent topics, I led a workshop introducing faculty and staff to Agentic AI on May 21, 2026. Below are the slides from my workshop. The workshop was prepared and delivered as part of my role as a Faculty Fellow at the Dominguez Center for Data Science.

Workshop – Beyond Prompting – A Gentle Introduction to Agentic AI, Presented at Bucknell University, May 21, 2026. ©2026 by Brian R. King, licensed under CC BY-NC-SA 4.0. To view a copy of this license, visit https://creativecommons.org/licenses/by-nc-sa/4.0/

Back to top ]]> AI GenAI Workshop https://www.eg.bucknell.edu/~brk009/posts/2026-06-09-workshop-beyond-prompting-a-gentle-introduction-to-agentic-ai-may-21-2026/ Tue, 09 Jun 2026 04:00:00 GMT arXiv: Preprints citing preprints https://www.eg.bucknell.edu/~brk009/posts/2026-05-23-arxiv-preprints-citing-preprints/ I’ve been reading a lot of AI research lately. Something has been nagging at me: the modern AI research ecosystem has quietly abandoned peer review, and the field has largely stopped noticing. I’m wrestling with understanding what this means, why it bothers me, and maybe recognizing I need to adjust my expectations in these times of incredible, rapid AI-assisted research on AI and AI-related topics.

A Brief History of arXiv

arXiv (pronounced “archive”) launched in 1991 as a preprint server for physicists who wanted to share results quickly before the long formal publication process concluded. The idea was simple and good: don’t lock up knowledge for 18 months behind a slow editorial queue. Let researchers read and build on each other’s work in near real-time.

It worked beautifully for physics. It has since become the dominant dissemination channel for computer science, mathematics, and AI. And in principle, that’s still a good thing. We now have a venue for research that is open, fast, and accessible to everyone regardless of institutional subscription status.

Is this OK? Or has something gone sideways?

The Citation Chain Problem

Here’s the pattern I keep running into: I open a recent AI preprint (something on large language model reasoning, reinforcement learning, or model training) and I look at its reference list. Nearly every citation reads something like:

arXiv preprint arXiv:2501.XXXXX

Not “published in NeurIPS 2025.” Not “Journal of Machine Learning Research, Vol. 22.” Just: preprint.

Some of the most widely cited papers in AI right now are unreviewed technical reports. The Qwen2.5 technical report (Qwen Team, 2024), describing one of the most widely used open-weights model families, is an arXiv preprint with (checking Google Scholar) has over 11,000 citations at the time of this post! The Qwen3 technical report (Yang et al., 2025)? Also a preprint. A paper I’ve recently been studying with my students, TTRL -a paper on test-time reinforcement learning that entire lines of follow-up work have been building on (Zuo et al., 2025) – initially circulated as a preprint before being accepted to NeurIPS 2025, meaning it spent months in heavy citation before clearing any formal bar. The NeurIPS reference has a little over 200 citations already. With respect to the utility of arXiv, as a Mandalorian would put it, “This is the way.”

And then there’s the category that doesn’t even seek conference review: industry technical reports from Meta, Google, Alibaba, and DeepSeek, describing frontier model families. The Llama 3 technical report from Meta (Grattafiori et al., 2024) is an arXiv preprint. DeepSeek-R1 (Guo et al., 2025) – the paper on incentivizing reasoning in large language models via reinforcement learning – circulated as a preprint for months before eventually being published in Nature in September 2025. That eventual publication is worth noting: the community built heavily on it, launched dozens of follow-up preprints, and treated it as established science well before any independent reviewer had looked at it. The formal peer review was really a lagging footnote to a citation trail that was already quite deep and impressive.

What does this means in practice for the AI researcher today? A new paper cites 30 sources, many of which were unreviewed at the time of citation. That new paper is itself a preprint. And within weeks, other preprints will cite it. The epistemic dependency chain is unvalidated all the way down. Should this be of concern?

Where Peer Review Has Gone

Traditional peer review assumed a sequential process: experiment → submission → expert review → revise → publish → cite. That model has effectively collapsed in AI/ML for a few compounding reasons:

  • Speed asymmetry. Peer review takes 6–18 months. The field moves in weeks. By the time a paper clears formal review, it may already have dozens of preprint descendants.
  • Venue congestion. Top conferences like NeurIPS, ICML, and ICLR, to name a few, accept roughly 20–25% of submissions at best. This means a large fraction of legitimate, solid work never clears the bar, not because it’s wrong, but because the venues are already overwhelmed.
  • Industry reports bypassing the queue entirely. Many of the most-cited papers in AI right now aren’t rejected conference submissions languishing on arXiv — they’re technical reports from industrial labs that were never intended for peer-reviewed venues. They describe systems their authors built and deployed, not experiments submitted for external scrutiny. This is a different kind of problem from slow review: it’s the deliberate absence of any review at all.
  • Thin conference review. Even papers that do get accepted at major venues typically receive 2–3 reviews, often written in 2–3 weeks by reviewers who may themselves be preprint-only researchers working in adjacent subfields. Having been a reviewer, it is a time-consuming process. The rewards are intrinsic, of course. It forces me to keep up to date, but in a field like AI today that is inundated with research, the process is too slow. arXiv is not the adversarial, months-long scrutiny that characterizes review in medicine or biology.

The result: the word “preprint” has functionally lost its meaning as a cautionary label. In AI, is a preprint now considered a de facto publication? It surely seems to be the case. It gets cited like one, benchmarked against like one, and built upon like one, often before anyone outside the authors’ own institution has carefully checked the work.

Citation Laundering

There’s a deeper problem I’d call citation laundering: a claim gets repeated across enough preprints that it acquires the social authority of an established fact without ever acquiring the epistemic warrant.

Consider how this plays out in practice. Paper A (a preprint) reports substantial accuracy gains on a reasoning benchmark using a new training method. Papers B, C, D, and E (all preprints) each cite Paper A as a foundational result and build refinements on top of it. Paper F then cites B through E, and its introduction reads as though A’s result is settled science. At no point in this chain has anyone outside the original research group independently verified the foundational claim. If Paper A overstates its result (through cherry-picked benchmarks, a subtle flaw in experimental design, or a training setup that doesn’t generalize at all), then all of B through F inherit that flaw.

This is not a hypothetical. There is a growing body of work raising exactly these concerns. Papers like “No Free Lunch: Rethinking Internal Feedback for LLM Reasoning” (Zhang et al., 2025) and “How Far Can Unsupervised RLVR Scale LLM Training?” (He et al., 2026) examine specific training paradigms — methods that use internal model signals instead of external reward supervision — and show that the gains they report tend to follow a rise-then-fall pattern: performance improves early in training, then collapses below the pre-training baseline. Aggregate benchmark numbers, looked at before that collapse sets in, would appear to be a genuine advance. Notably, both of those critical papers are also arXiv preprints.

The specific findings in those papers are about a particular class of methods, not a sweeping indictment of all AI benchmarking. But they illustrate the general risk well: a training approach can produce results that look compelling at the wrong moment in its training curve, get cited heavily in that window, and the problematic training dynamics surface only later when someone looks harder.

Perhaps this is the new form of peer review in rapidly evolving fields such as AI: release a preprint on arXiv, let it get some press, make the rounds on the socials where people (including the authors themselves) can post and hype up their work, or much more preferably, start discussed on review sites (e.g. https://gotit.pub) until another preprint comes out to critique and improve on prior work.

The Case for arXiv (Being Fair)

I don’t want to be entirely one-sided here, because the alternative, i.e. returning to traditional journal timelines, is not better. In fact, I could argue it is not suitable for AI at this time (though I would disagree with myself on that latter point!)

Peer review has its own well-documented failure modes. It is slow, biased toward established labs and prestigious institutions, and has failed catastrophically to catch replication crises in fields like social psychology and biomedical research that used it faithfully. Peer review is a filter, not a guarantee.

The open-access aspect of arXiv is also genuinely democratizing. Researchers at institutions without expensive journal subscriptions can fully participate in the conversation. That matters.

And the AI community does exercise a form of informal community review. As I mentioned above, it’s pretty safe to say that preprints risk getting publicly scrutinized on social media. Competing labs will work to replicate (or fail to replicate) the work, and will be challenged by follow-up work. For code and results that can be independently reproduced, this is actually quite fast and sometimes more effective than formal review. DeepSeek-R1 is again instructive: within days of its preprint release, multiple groups were attempting to reproduce its results, flag discrepancies, and extend its methods. The stronger the hype behind a paper (remember that huge dive NVDA took after DeepSeek?!?) So, that review process is real, and seems to work… most of the time. It just happens in public, messily, over months, rather than privately before publication.

The limitation of that informal process is also real, though: it works well for claims that can be reproduced from public code and standard benchmarks. It performs poorly for claims that hinge on proprietary data, undisclosed training details, or subtle methodological choices not visible in the paper. Those claims can circulate unchallenged for a long time.

The Problem Is the Conflation

Here’s my actual concern, stated plainly: arXiv is enormously useful as a communication tool, but it has been mistaken for a validation tool.

When a paper citing dozens of unreviewed technical reports is itself cited as authoritative in another preprint, we are building an increasingly tall structure on an unverified foundation. For fast-moving engineering claims that get stress-tested by replication at competing labs, this is uncomfortable but arguably tolerable. For deeper scientific claims about what these models actually learn, how they generalize, and whether reported gains reflect genuine capability acquisition, it is a meaningful epistemic risk that the field has largely chosen to accept without much reflection.

What would be better? At minimum: more careful hedging when citing unreviewed work, especially in introductions and related-work sections where preprints often get laundered into settled background. Journals and conference proceedings could normalize explicitly flagging which cited works were unreviewed at time of publication. Science journalists covering AI results could routinely note preprint status the way they note funding sources. None of these are radical changes in the process. It just requires some agility.

I know, I know. Agility is to academia as oil is to water.

Regardless, the speed of AI research is genuinely exciting. But speed without validation is just noise that moves fast.

It’s a reminder for all of us to be cautious and critical of all new preprints. Maybe we should have been doing this with all new research anyway.

Back to top ]]> AI Musings Research https://www.eg.bucknell.edu/~brk009/posts/2026-05-23-arxiv-preprints-citing-preprints/ Sat, 23 May 2026 04:00:00 GMT Conscience Is Not One-Sided https://www.eg.bucknell.edu/~brk009/posts/2026-05-12-conscience-is-not-one-sided/ I have a lot of respect for Alex Reid. He has a lot of great posts that often help me bring some balance with AI and technology in general. His recent post, “The Conscience of AI Refusal”, raises questions that matter deeply to those of us who spend our professional lives thinking about how, what, and why we teach. Reid draws on a recent resolution passed at the Conference for College Composition and Communication, affirming the right of instructors and students to refuse generative AI in the writing classroom. He built a thoughtful philosophical argument grounding that refusal in conscience. It’s worth a read. I have genuine respect for that argument. But I think it stops one step short of where we actually need to go, and that gap has real consequences for higher education.

I am not here to dismiss conscientious refusal, defend corporate imposition of AI tools on unwilling faculty, or pretend that AI is somehow culturally neutral or free from the market pressures (which Reid rightly identifies). On all of those points, he and I agree completely. No institution should be demanding that all faculty use AI or any other tools, for that matter, if it violates their conscience. Besides, there is that thing called academic freedom (though some may argue elements of that are eroding away, too.)

What I want to challenge is this: the implicit assumption that conscience, in this debate, belongs primarily to one side. It absolutely does not, and should not be construed as such.

Conscience Has Two Edges

Reid argues that if we accept AI as culturally and historically embedded, it becomes “unconscionable” to view it simultaneously as an efficient cognitive proxy. I understand the logic, but I’m not sure it holds here. There is no lack of examples one could consider:

  • The printing press transformed power structures, enabled propaganda and religious divides (and still does), was (and still is) deeply political and (still is) commercially driven, but also genuinely democratized literacy. No one seriously argues we should have refused it on cultural grounds alone.
  • The calculator. Ah, one of my favorite examples. Didn’t math educators have the exact same debate many decades ago, suggesting that using it erodes deep numerical reasoning? It’s culturally situated (built by corporations, shaped by market forces) and it frees cognitive load for higher-order thinking. Both things are true. (If mathematicians think AI isn’t going to affect them, think again.)
  • The textbook is a commercial artifact produced by publishers. It has enormous economic and ideological interests, yet we don’t conclude that using textbooks is epistemically unconscionable.
  • The internet is probably the best example here. It is culturally saturated, corporately controlled, surveilled, and politically weaponized, and used for all sorts of horrific things (social media for starters!) And yes, it’s also the medium through which experts like Reid and I publish our thoughts for the community.
  • Writing itself is one you’ll see over and over in this debate. I was reminded by a friend earlier this year how Plato famously argued that writing would weaken memory and corrupt genuine knowledge! He was not wrong about writing’s cultural embeddedness. He was wrong to conclude that refusal was therefore the conscientious response.

Acknowledging the cultural situatedness of a tool does not, by itself, settle the question of whether or how to use it. More importantly, integration, i.e. the thoughtful, critical, pedagogically intentional use of AI, is itself available as a conscientious act. I’ll say it – I am deliberately bringing AI into my classroom, and I do so because I believe, in good conscience, that I have a responsibility to prepare students for a world already being reshaped by these tools. To send them out without the deep experiences to learn how to use AI in responsible ways, and likewise to critically interrogate and challenge AI intelligently, is, to me, its own form of abdication. That is my conscience speaking. It is no less grounded in shared knowledge than the conscience of refusal, right?

The Symmetry We Must Name

The framework must be applied symmetrically. If refusal can be an act of conscience, then can’t engagement be as well? The moral seriousness of the act is not a property of the direction you choose. It is a property of the care, honesty, and accountability with which you choose it. To deny that symmetry is to do exactly what Reid cautions against in his closing lines: to become a “refuser of the refusers,” generating an infinite regress of competing moral condemnations. He sees this danger clearly when it threatens the refusers. We need to see it equally clearly when it threatens those who integrate.

The Real Threat to Higher Education

Higher education is already in crisis. Trust in institutions is eroding. The political and cultural fractures running through our society run straight through our campuses. Bucknell is not unique in this challenge. The impact will be the same, no matter what type of institution you are. In this environment, nothing accelerates our decline faster than faculty turning on each other over questions of pedagogical conscience. That dynamic is not hypothetical – it is already playing out in our departments. If the dominant message coming out of disciplinary organizations is that one side of this debate has conscience and the other does not, we will drive a deeper wedge into an already fractured community at precisely the moment when students need us to model something better. What we need is not consensus. We need healthy discourse with rational arguments. Not feelings and emotions. Genuine disagreement is healthy. It’s good for any relationship! What we need is mutual recognition: the shared acknowledgment that reasonable, thoughtful, ethically serious educators can look at the same situation and reach different conclusions.

What Good Conscience Actually Requires

Stengers, whom Reid invokes to powerful effect, urges us to “think with” our tradition rather than transcending it through withdrawal. For Stengers, this is not a passive move, but a demand to stay inside the friction and resist the temptation of a clean exit. The AI-refuser still inhabits a world saturated with AI-generated text, AI-assisted research, and is likely using a profound number of tools built on AI (e.g. autocorrect, spell check, predictive text, spam filters, Google Search, navigation with Google Maps, product recommendations with Amazon or Netflix, or fraud detection for questionable credit card purchases, to name a few.) More importantly, the refuser teaches students who will enter AI-shaped workplaces. And likewise, the AI-adopter and AI-integrator will need to carry the weight of what these tools displace, distort, and commodify. Both positions are forms of dwelling in the difficulty. The question is only how we dwell there, and with what level of honesty about the costs. Neither side is clean. Neither is innocent. Both are plagued with challenges moving ahead. What neither side can afford is to treat its own conscience as the universal standard against which the other is measured and found wanting. That is not conscience! That is a mirror being held up as a window – one’s own reflection being mistaken for an objective view.

Here is what I am asking of all of us: Let’s build an academic culture where we can look across the aisle at a colleague who has made a different pedagogical choice and say, “I see what you are doing. I may not make that choice myself. But I recognize that you are doing it seriously, thoughtfully, and in good faith.” And then let’s get back to the shared work of teaching students well, in a world neither of us fully controls. To me, our students need a path that lets them experience both sides of the AI divide and develop the ability to follow their own conscience.

That is the conscience higher education needs right now. And it belongs to both sides.

Back to top ]]> AI Musings https://www.eg.bucknell.edu/~brk009/posts/2026-05-12-conscience-is-not-one-sided/ Tue, 12 May 2026 04:00:00 GMT Is Computer Science Dead? https://www.eg.bucknell.edu/~brk009/posts/2026-05-02-is-computer-science-dead/ There is very little doubt in my mind that AI is here. Despite the plethora of thoughts and feelings of faculty across higher ed, AI is here, with all signs pointing to its continued growth and adoption. Its capabilities are growing at an absolutely profound rate, making any effort to try and figure out the best approach seem futile at best. It has introduced a level of disruption unlike anything we’ve seen, and nowhere is this more true than in computer science.

I have had quite a few discussions over the past couple of years about AI. The topic is often around how we are preparing our students. The discussions have involved every direct and indirect member of the Bucknell community, including prospective students and their parents, current students, our alumni, companies that consider hiring our students, staff on campus, and members of our administration at all levels, including the provost’s office, admissions, and university advancement. I also recently spoke before our Board of Trustees (as recently as last week) about how we are working to prepare our students. The only common takeaway from all of these conversations is that we’re all concerned, and we’re all working in our own way to provide the most meaningful path forward for our students.

I have been restructuring my courses to incorporate hands-on experience with AI tools, and I am far from alone in this effort. Many of my colleagues are doing the same. Across all three of our colleges, faculty are working diligently to find meaningful ways to integrate the thoughtful, careful, and responsible use of AI into their courses without encouraging cognitive offloading, while still giving students productive pathways for using these tools effectively. I believe we have a responsibility to equip all students, regardless of major, with strong AI literacy skills so they can thrive in an evolving workforce. However, there is no one uniform approach for all majors. The impact of AI across disciplines is wide and varied. The rapid evolution of AI technology sometimes leaves one feeling hopeless in education. It seems like the moment we convince ourselves we may have solid ideas moving forward, they quickly become irrelevant as new capabilities arise. Indeed, these are exciting times!

Computer science is being hit hard, but we have survived downturns repeatedly in the past, and we will survive this. However, this downturn is fundamentally different. The core tools of our trade are no longer relevant! It’s time for computer scientists and the discipline itself to evolve. We must recognize that we are in the midst of a new industrial revolution in our field. Everything we’ve been doing up to this point needs substantial renovation and reimagination. And, let’s agree that the field of Artificial Intelligence itself ought not to be considered only the domain of computer science. Sure, it rose in popularity largely because of substantial advances in computer science, computer engineering, electrical engineering (all brought to you our two lovely parents, math and physics!) Today, AI is not just a computer science or STEM initiative; it is a campuswide responsibility, one where the humanities and social sciences must play an essential role. How can it not? How can we teach artificial intelligence without having students in AI also learn about the very systems these computational algorithms are designed to replicate, i.e. human intelligence. It’d seem prudent that, at a minimum, any core AI curricular endeavor ought to include content that delves into human cognition.

In this new landscape, the urgent task of computer science education is not to outcompete AI at writing code. I cannot think of a more futile task. Writing code is no longer the exclusive marker of expertise and computational prowess it once was. (And for those like me who grew up thriving as teenage nerds writing our own games, and programs to solve our math problems in high school – just because we could – can I just say… damn, that sucks! I loved coding!) AI tools today can perform much of that work far more quickly, effectively, and efficiently than any programmer could.

So, is computer science dead? Absolutely not! We need to evolve, and those who refuse to evolve risk becoming irrelevant. We must cultivate and integrate the human capacities that AI cannot replace. Our students must learn to think critically and independently, to define and understand meaningful problems, and to evaluate AI-generated solutions for correctness, complexity, efficiency, and fitness to real-world constraints. They need experience working with people, assessing real-world problems through client communication, and defining their problems with solid goals and constraints that translate to plans that can easily be carried out in tandem with AI. They need to reason about consequences when systems fail or behave unpredictably, because AI, like humans, will never produce perfect solutions, especially as long as humans are orchestrating them. They must be able to communicate clearly with both technical and nontechnical audiences, design for people by optimizing for user experience and considering accessibility needs, and collaborate across disciplines to understand the social and organizational contexts in which their systems will operate. And ultimately, they must develop the ethical judgment to decide when and how a system should be built, deployed, or rejected altogether. These are not peripheral skills; they are the core of what it means to be a computing professional in an age where AI can generate code but cannot assume responsibility for its impact.

Computer Science Skills in the AI Era

It’s worth repeating. Let’s summarize some of the skills we want to be injecting in our courses throughout our curricula in CS. To be honest, I do not see anything new in my list. In fact, one could argue that AI is an amazing realization that is finally allowing us to put emphasis on things that computer science should have been emphasizing all along:

Critical thinking and independent reasoning – evaluating AI outputs rather than accepting them uncritically

Problem definition and framing – identifying what question is actually worth solving

Solution evaluation – assessing correctness, algorithmic complexity, efficiency, and real-world fitness

Consequence reasoning – anticipating failure modes, unintended behavior, and downstream effects

Communication – speaking and writing clearly for both technical and nontechnical audiences

Collaboration and teamwork – leading and contributing in interdisciplinary groups

UX and human-centered design – designing for real users, not just functional outputs

Accessibility and inclusion – building systems that serve diverse populations from the start

Debugging and system sense-making — diagnosing complex and AI-assisted systems rather than treating outputs as oracles

Data literacy – understanding data quality, bias, and the limits of models trained on it

Metacognition – knowing when AI is likely wrong and when to slow down or seek other perspectives, step back, and reconsider your plan

Adaptability and lifelong learning – keeping pace with rapidly evolving tools and ecosystems

Domain reasoning – applying computing thoughtfully within specific real-world fields and constraints

Ethical judgment – reasoning about fairness, responsibility, privacy, power, and accountability

Inspiring AI Literacy Initiatives

I’ll close with a handful of examples I’ve been collecting of how colleges and universities around the country are embracing AI with a liberal arts mindset:

  • Why AI is ‘resurrecting’ the liberal arts for the Class of 2026: At Wake Forest, educators are observing that the rapid evolution of workplace technologies is actually increasing the value of traditional liberal arts strengths. Rather than making humanistic study obsolete, AI platforms require graduates who can critically evaluate outputs, communicate effectively, and apply broad context to complex problems.
  • Building the Future: Why Teaching AI to Liberal Arts Students Is Critical Work: Dartmouth’s Tuck School of Business emphasizes that liberal arts students possess the exact intellectual framework necessary for an AI-augmented workforce. Their initiative demonstrates that non-STEM majors can excel in tech-driven environments when taught to leverage their natural abilities to ask probing questions and maintain ethical oversight.
  • Social scientists embrace the AI moment: Stanford researchers are highlighting how AI is fundamentally transforming empirical workflows and data analysis within the social sciences. By integrating AI into these fields, students learn to navigate new research methodologies while applying human judgment to automated text analysis and summarization tools.
  • Generative AI is raising new questions for liberal arts education: The University of Richmond is actively confronting the ethical and pedagogical challenges brought by generative models by fostering cross-disciplinary discussions. They are creating practical campus resources that help faculty and students collaborate to critically engage with AI tools across all departments.
  • Breaking Faculty Barriers to AI Literacy: The Digital Education Council provides a measured look at the widespread institutional struggle to move from mere interest in AI to actual classroom implementation. They argue that addressing faculty hesitation requires dedicated support structures, clear incentives, and practical guidance rather than just top-down mandates.
Back to top ]]> AI Musings computer science curriculum https://www.eg.bucknell.edu/~brk009/posts/2026-05-02-is-computer-science-dead/ Sat, 02 May 2026 04:00:00 GMT AI Agents should not be used as customer service https://www.eg.bucknell.edu/~brk009/posts/2026-03-05-ai-agents-should-not-be-used-as-customer-service/ We’re hearing that multiple companies are striving to optimize and streamline services across their businesses with AI, buying into the promises AI companies themselves are touting. However, this is all too new to really understand best practices for AI deployment. Beyond software development, we have very little evidence to justify the rate of adoption we’re seeing. Sadly, so many are jumping on the AI bandwagon out of FOMO – fear of missing out. They are not taking the time to understand how the blind deployment of these agents in the wrong way may have long-lasting detriments, many of which will be hard to recover from. It’s a misguided optimization that ignores the customer.

Here’s the latest example – deployment of AI agents as the front line for customer service.

Quite a sobering read, really. It’s a pathetic, absurd move on the part of these businesses that think that deploying AI agents as customer service is going to win over customers. But, don’t blame AI! Blame the AI vendors:

The widespread adoption of obstructive AI is not just a corporate decision; it is fueled by a booming industry of AI vendors that market their platforms with promises of massive cost savings. An analysis of marketing materials from leading vendors like Talkdesk, Intercom, Zendesk, and Five9 reveals that the primary value proposition is not better service, but radical efficiency achieved by reducing human involvement.

As an AI/ML researcher, let me explain what’s happening. AI models are algorithms that learn from data. They can be trained to optimize for whatever problem the AI engineer is tasked with solving using these powerful tools. In machine learning, we frame learning as an objective function – a simple, measurable quantity that we want the model to learn to minimize or maximize. It’s not clear what these companies are optimizing for. Any number of possible metrics are possible here.

  • Minimize escalation rate – reward the model for closed tickets that do not involve humans
  • Minimize average handling time, regardless of whether the customer is satisfied
  • Minimize refund approval rate – this is pretty diabolical – let the model learn patterns in dialogue that fatigue the customer the most and lead to them not pursuing any refund!

This is not the fault of AI, but represents a deliberate design failure. AI companies will train models with whatever objectives they choose. They certainly have no reason to do otherwise. And when you fine-tune an AI agent to minimize escalation that involves humans, it will learn all sorts of delay tactics. Repeating responses, asking for clarification, asking for documents already submitted, offering irrelevant options, etc. There are many ways it can learn to fatigue a human! 

A well-designed AI agent should be optimizing for customer-side outcomes. Why aren’t these companies optimizing for this value proposition? (I smell greed!) For example, maximizing customer resolution rates or minimizing downstream complaints being filed are just a couple. The moment the company’s cost function and the customer’s satisfaction function are treated as the same optimization target, the entire problem changes, and these AI agents would learn to improve their dialogues in ways that leave both the customer and the company happy. But that’s not where companies are at right now. Businesses are blindly jumping on the AI bandwagon.

Finally, it’s worth reminding all of us that when we’re engaging with AI models, we’re engaging with models that learned how to respond with streams of tokens, i.e. words that represent the most likely responses of words that optimize the AI model’s objective function. There is absolutely no affective intelligence – the capacity to perceive, interpret, and respond appropriately to human emotional states. You and I do this automatically. We’re pretty awesome that way. We can detect frustration in word choices. Hesitations and pauses are meaningful. The phrase, “your silence is deafening,” would be meaningless for an AI agent. We can hear distress in tone. We can adapt and adjust in real time (or not!) Even the most sophisticated models today, at best, can only predict a particular emotional state based on selected words, but can’t infer the meaning of that emotion in context. While this is an ongoing area of research, tonality in human voice is so wide and varied that it’s indeed a challenging problem for AI. None of these deployed agents are designed to properly channel empathy, moral intuition, compassion, dignity, respect, etc. Of course, a bigger question is, would we even want that? Personally, I would not. A compassionate AI agent is still just an algorithm.

I can’t help but wonder how long it will be before this practice reflects negatively on companies that deploy AI agents as their first-pass interface for their customer base. As an AI researcher who teaches students how to build these models, I find this (among a variety of other misguided uses of AI today) quite unfortunate; these choices only deepen the public’s distrust and growing weariness of AI. 

So, here’s your PSA: Recent surveys and research show that humans have a strong preference not to interact with AI agents. (And, if you’ve been subject to interacting with an AI agent when you called for customer service, you get it!) For example, one recent online survey of 1,011 U.S. consumers found that 93.4% prefer interacting with a human over an AI agent. Another study suggested 70% of customers will abandon your brand after just one poor AI experience. It makes sense – if a company thinks so little of their customers that they are willing to make us talk to computers instead of humans, people will not remain loyal.

My most recent personal interaction with an AI agent is partly what drove this post. I had to place a call to my local healthcare provider recently. Even in healthcare, they are replacing their front desk receptionists with AI agents. The whole experience left me wanting to search for an alternative provider. I ultimately ended up needing a human agent anyway, who could have simply resolved my question in a small fraction of the time it took with the AI agent.

Let’s hope some of these companies learn to put humans back as the primary interface for the company and keep AI agents focused on what they are good at. If they don’t, I see a big opportunity for the smaller, local businesses to see a resurgence in many domains. Humans do not want to interact with AI! Why is this hard for businesses to understand? (Greed blinds common sense.)

The big Gartner report last month gives me some hope – they are predicting that half of the companies that laid off customer service staff will be forced to rehire by 2027. 

AI is a tool to assist humans in their work. It should not replace humans, especially in human-facing roles. Companies that choose to use it this way will experience quite a backlash in business.

Back to top ]]> AI Musings https://www.eg.bucknell.edu/~brk009/posts/2026-03-05-ai-agents-should-not-be-used-as-customer-service/ Thu, 05 Mar 2026 05:00:00 GMT Talk – Prompting AI – July 15, 2025 https://www.eg.bucknell.edu/~brk009/posts/2025-07-15-talk-prompting-ai/ I gave a talk for faculty and staff on how to prompt AI to get better results. We covered the background of AI, explained LLMs and their data, and discussed the increasingly common caveats and societal concerns with blindly using chatbots like ChatGPT, Claude, etc. We then covered the essentials of prompt engineering, with beginner and advanced tips.

The talk was mostly focused on summarizing, investigation, and brainstorming tasks, not on improving day-to-day mundane tasks such as automated e-mail handling and such. (That will be a future talk!)

Key slides related to the prompting part of the talk are linked below:

Slides – Prompting AI for Faculty/Staff – July 15, 2025

These materials are part of workshops and talks I give as part of my role as Faculty Fellow of the Dominguez Center for Data Science.

Prompt Engineering, Presented to Faculty/Staff at Bucknell University, July 15, 2025. ©2025 by Brian R. King, licensed under CC BY-NC-SA 4.0. To view a copy of this license, visit https://creativecommons.org/licenses/by-nc-sa/4.0/

Back to top ]]> AI Talk https://www.eg.bucknell.edu/~brk009/posts/2025-07-15-talk-prompting-ai/ Tue, 15 Jul 2025 04:00:00 GMT AI coding tools make developers slower but they think they’re faster, study finds https://www.eg.bucknell.edu/~brk009/posts/2025-07-14-ai-coding-tools-make-developers-slower-but-they-think-theyre-faster-study-finds/ https://www.theregister.com/2025/07/11/ai_code_tools_slow_down/

Download the full paper from metr.org. 

I want this paper to join the regular stream of AI articles appearing in everybody’s news feed, especially those of the CEOs of Anthropic, OpenAI, Meta, Google, Microsoft, and every other company out there banking on being able to downsize their workforce someday on the promise of the AI hype. As is so often the case in the tech industry, the CEOs are out of touch with reality.

The setup

The study was a randomized controlled trial (RCT) involving 16 experienced open-source developers, each with an average of 5 years of prior experience on mature projects. Developers completed 246 tasks – Github issues to be more specific – with each task randomly assigned to either allow or disallow the use of AI tools. The AI tools they allowed were Cursor Pro, configured with Claude 3.5/3.7 Sonnet LLM.  When AI tools were allowed, developers primarily used Cursor Pro and Claude 3.5/3.7 Sonnet. The study included onboarding and training, required developers to record their screens while working, and collected both quantitative and qualitative data through surveys, interviews, and screen recordings. They tried to mirror real-world development closely. 

The results

Developers were quite confident in AI before the study. On average, they thought they woudl shave 24% off their development time. In contrast, they ended up increasing their development time by 19%! 

They postulated about a lot of factors that could contribute to the slowdown. With such a small population in this study, it’d be hard to find a lot of strong predictors. However, there were some that met significance that were quite telling:

  • Over-optimism about AI usefulness – On average, the developers predicted their implementation time would decrease by 24%. Instead, it increased by 19%.
  • High developer familiarity with repositories – This confused me at first, but as a long-time developer, it made sense. Developers who were intimately familiar with their repositories should trust themselves to work on their issues rather than trust AI. AI will be challenged to match human expertise on a project in which you are already the expert. 
  • Large and complex repositories – This should come as no surprise. Some of the repositories that had the worst outcomes were those with over 1,000,000 lines of code. No human or AI is going to be able to easily make sense of that, especially when the repo has been around a long time and is riddled with legacy, outdated, poorly maintained code. What makes people think that AI is going to somehow do a better job of fixing crap code? It can help a developer refactor it so that it can then meet high standards, and then maybe it’d have a better chance of helping with improving and fixing the code. 
  • Low AI Reliability – Most developers (myself included) regularly deal with AI generating useless code. But it’s confident about its useless output! So convincing! Thus, devs often erroneously just accept it. Unfortunately, we eventually get into a loop where we accept buggy code, ask AI to fix its own buggy code, at which point it introduces another bug. Sometimes, it’ll even go back and re-introduce the bug it originally fixed. It’s a vicious cycle.

I would strongly advise any company considering downsizing its developer workforce to carefully reconsider their plans. AI has its strengths. The developer who blindly uses it with no real knowledge ot software engineering is an incredibly dangerous developer who should be avoided. However, a developer who knows AI’s strengths and weaknesses, who knows when to trust it, when to avoid it, and how to control and leverage it carefully in thier workflow has tremendous promise. They will be in sought-after positions for quite some time.

Back to top ]]> AI GenAI Software Engineering https://www.eg.bucknell.edu/~brk009/posts/2025-07-14-ai-coding-tools-make-developers-slower-but-they-think-theyre-faster-study-finds/ Mon, 14 Jul 2025 04:00:00 GMT How Do You Teach Computer Science in the A.I. Era? – NYTimes https://www.eg.bucknell.edu/~brk009/posts/2025-07-01-how-do-you-teach-computer-science-in-the-a-i-era-nytimes/ https://www.nytimes.com/2025/06/30/technology/computer-science-education-ai.html

Key takeaways for the CS professor:

Computer Science as a field of study has been shaken to the core. It’s not that it’s no longer relevant. In fact, it’s probably more relevant now than ever – as long as you’re willing to broaden and reconsider what it even means to be a computer scientist in the AI era. The best schools out there, such as Carnegie Mellon, are reevaluating their curricula in response to generative AI tools like ChatGPT. This is something we’re actively in the midst of doing here at Bucknell. I can’t imagine any computer science program today remaining relevant unless they consider a massive overhaul.

The fact is that AI is rapidly transforming how computer science is taught, especially coding. Traditional programming should no longer be considered a primary objective in any CS curriculum. We are in the midst of transforming our curricula to consider broad topics such as computational thinking and AI literacy.

Ideas to consider:

  • Computer science may need to evolve toward a liberal arts-style education. We’ll need to consider more interdisciplinary and hybrid courses that integrate computing and computational thinking into other disciplines.
  • More courses and experiential learning opportunities that focus on critical thinking, ethical use of AI, and communication skills. I would also argue this is prime time for computer science programs to finally put heavy emphasis on User Experience – something that AI is horrible at. Any aspect of our field that focuses on the human side of our product is essential. UX, Human-computer interaction (HCI), user interface design, teamwork, project management, communication and presentation skills, data visualization, and so on, need to be incorporated in more courses. We no longer have the excuse of not having the space in our program to cover these essential skills.
  • Early computer science courses still need to stress computational thinking. AI will get 90% of the job completed, but will continue to struggle with the most complex pieces of large-scale projects. The more complex the project, the more it will struggle. Unfortunately, students often have a false sense of security and confidence, blindly using AI with no knowledge of how to fix problems they find, or even worse, they’ll lack knowledge of how to properly debug and test systems for correctness in the first place, operating under the dangerous assumption that the AI generated code is correct.

Career Impacts

I think we can all agree that AI seems to be eliminating some entry-level coding work, though I am struggling to get any real numbers on how AI is impacting these jobs vs. economic factors. The job market has tightened—entry-level roles are fewer, and students are applying more broadly. But here’s the complete story, and one that this NYTimes article concludes as well. I’ve been telling prospective and current computer science students that the news narrative is not telling the whole story. Based on where our students are still getting jobs today, AI is taking the jobs of those who do not know how to leverage AI. That’s a pretty clear, widely accepted reality in our field. Here’s the kicker that is rarely being reported (because, again, the news cycle thrives on negativity) – despite layoffs, demand for AI-assisted software is growing!If you have AI literacy skills, combined with strong, demonstrable critical-thinking skills, you can share experiences in and (preferably) outside of the classroom that show you know how to orchestrate solutions to large-scale projects, you can work well with teams and communicate and present results, and for goodness sake you understand human-centered design and how to measure and maximize UX, you will continue to remain in a highly sought-after field!

My final thought from this article – It’s pretty clear that AI has not only democratized cheating, but with respect to computer science, it has democratized programming:

“The growth in software engineering jobs may decline, but the total number of people involved in programming will increase,” said Alex Aiken, a professor of computer science at Stanford.

More non-tech workers will build software using AI tools. It’s a slippery, dangerous slope. Why?

“But they didn’t understand half of what the code was,” he said, leading many to realize the value of knowing how to write and debug code themselves. “The students are resetting.”

That’s true for many computer science students embracing the new A.I. tools, with some reservations. They say they use A.I for building initial prototype programs, for checking for errors in code and as a digital tutor to answer questions. But they are reluctant to rely on it too much, fearing it dulls their computing acumen.

Indeed, this is the reality check that we’re seeing in our own students here. They notice that AI is not always correct, even when solving simpler undergraduate computer science exercises. Conclusion: students still need to understand the fundamentals of computer science to fix the convoluted sequence of tokens from the LLMs. The output generated will appear to be a solid solution at first, looking like well-written Python, Java, C, or whatever language you’re working in, even properly commented if you prompted it correctly. And heck, the chatbot will sound extraordinarily confident and cheeky about its solution, pleased to have served you! But… it’s still just a stochastic sequence of tokens, subject to error.

Not all that different than human solutions.

Back to top ]]> AI CS Education GenAI https://www.eg.bucknell.edu/~brk009/posts/2025-07-01-how-do-you-teach-computer-science-in-the-a-i-era-nytimes/ Tue, 01 Jul 2025 04:00:00 GMT The Entire Internet is Reverting to Beta – The Atlantic https://www.eg.bucknell.edu/~brk009/posts/2025-07-01-the-entire-internet-is-reverting-to-beta-the-atlantic/ https://www.theatlantic.com/technology/archive/2025/06/ai-janky-web/683228/

Nattei Wong has a fantastic write-up that is a bit of a different take on the societal impact of generative AI. Instead of yet another article on the dumbing down of humanity, instead, Wong alerts us to our increased tolerance of errors, mistakes, hallucinations, and general shoddy solutions that AI is generating. For some reason, though we have little tolerance of other humans making mistakes, we’re seemingly OK with AI making mistakes, deeming it all as acceptable.

The opening paragraph says it all: 

A car that accelerates instead of braking every once in a while is not ready for the road. A faucet that occasionally spits out boiling water instead of cold does not belong in your home. Working properly most of the time simply isn’t good enough for technologies that people are heavily reliant upon. And two and a half years after the launch of ChatGPT, generative AI is becoming such a technology.

His final take – AI is in a dangerous zone, noting, “They may not be wrong frequently enough to be jettisoned, but they also may not be wrong rarely enough to ever be fully trusted.”

The concern here is something we’re already seeing among the population using generative AI chatbots—we’re becoming “beta” as we become accustomed to and accept mediocrity. When it makes mistakes, we just say, “Oh yeah, it hallucinates. Oh well.”

Back to top ]]> AI GenAI ethics society https://www.eg.bucknell.edu/~brk009/posts/2025-07-01-the-entire-internet-is-reverting-to-beta-the-atlantic/ Tue, 01 Jul 2025 04:00:00 GMT Anthropic launches program to track fallout from AI https://www.eg.bucknell.edu/~brk009/posts/2025-06-27-anthropic-launches-program-to-track-fallout-from-ai/

As job losses loom, Anthropic launches program to track AI’s economic fallout

Here’s what you need to know:

Amid this backdrop, Anthropic on Friday launched its Economic Futures Program, a new initiative to support research on AI’s impacts on the labor market and global economy and to develop policy proposals to prepare for the shift. 

Anthropic, one of the big AI companies whose existence is entirely focused on the development of new AI tools, who recognizes they are pushing out AI tools with no concern of their impact on humanity, whose livelihood depends entirely on selling its product to greedy corporations worldwide, corporations who are buying into the hype of being able to save billions of dollars by displacing human skill, is investing in a new initiative to help understand how AI is impacting the labor market and global economy.

Let that sink in. Can anyone say “conflict of interest?”

Anthropic’s CEO, Dario Amodei, has been oddly vocal about their own destructive impact:

At least one prominent name has shared his views on the potential economic impact of AI: Anthropic’s CEO Dario Amodei. In May, Amodei predicted that AI could wipe out half of all entry-level white-collar jobs and spike unemployment to as high as 20% in the next one to five years. 

And yet they keep functioning, business as usual. I do not think anyone is so naive as to believe that any tech CEO has a genuine concern for how their product may (or may not) devastate humanity. Amodei has been a bit more extreme than the other CEOs in sounding alarms, and yet Anthropic has continued to release more capable AI models and tools on an unprecedented cycle.

Remember that time long ago (only a few years) when big corporations were expected to adhere to the triple bottom line?

The triple bottom line is a business concept that states firms should commit to measuring their social and environmental impact—in addition to their financial performance—rather than solely focusing on generating profit, or the standard “bottom line.”

Sadly, as is the usual outcome with humans, greed won, flipping the proverbial bird to human and societal concerns and the environment. The amount of harm these companies are actively doing to two of the three components suggests they have no actual concern for anything other than one bottom line – profit.

Suppose Anthropic, OpenAI, Perplexity, Tesla, Google, Amazon, etc., actually cared about the future of humanity and the planet. In that case, one might think they’d slow down, carefully evaluate the impact of their products and be more responsible. Perhaps they would provide massive funding for academics and researchers who are invested in this space to start working with them to study how to move forward in a way that is responsible to people and the planet.

They do not care. Virtue signaling through the announcement of some initiative to study the damage Anthropic and other AI companies are actively doing is a bit like the wolf draped in a gardener’s cloak, planting a couple seeds with one hand while uprooting the forest with the other.

Amodei (Anthropic), Cook (Apple), Altman (OpenAI), Zuckerberg (Meta), Huang (Nvidia), Nadella (Microsoft), Pichai (Google), Jassy (Amazon), and Musk (Tesla), are just a handful of CEOs who run companies highly invested in seeing AI “succeed” (which means, from their viewpoint, massive profits.) These companies are closing in on an aggregate value of $20 trillion dollars. The collective worth of the CEOs based on publicly available data could be around $350 billion dollars:

CEO Company Estimated Net Worth (USD)
Dario Amodei Anthropic $1.2 billion1
Sam Altman OpenAI $2.8 billion2
Mark Zuckerberg Meta $12.3 billion3
Jensen Huang Nvidia $135 billion4
Satya Nadella Microsoft $500 million5
Sundar Pichai Google $1.3 billion6
Andy Jassy Amazon $470–540 million789
Elon Musk Tesla $194 billion (2025 estimate)10
Tim Cook Apple ~$2 billion (2024 estimate, not in search results; based on public filings and recent news)

The above are estimates as of June 2025 and fluctuate with the market. They also represent an unfathomable amount of money that could be used to fix the problems they are creating. Ironically, the AI tools themselves might be an important part of developing solutions.

It’s time these companies do more to invest in education, science, and research so that we, as students and academics, can help them change the course of this behemoth of a ship and steer them toward responsible AI – an AI that does not neglect humanity and the planet. These companies all reside in a country that is actively working to dismantle education, research and science – all things that brought America greatness long ago. Do these companies care about the longevity of their product? Do they want to ensure they have a workforce in the future that can continue to contribute to the science and technology necessary for their own growth? These companies must pull their heads out of their huge profit mounds and use their massive pools of money to invest in our future. They should be doing far more than just releasing tools that let us investigate the obvious negative impacts they are having on the workforce and the economy.

Back to top ]]> AI GenAI ethics https://www.eg.bucknell.edu/~brk009/posts/2025-06-27-anthropic-launches-program-to-track-fallout-from-ai/ Fri, 27 Jun 2025 04:00:00 GMT AI as the New Study Partner: Yes, we CAN help students learn responsible AI use https://www.eg.bucknell.edu/~brk009/posts/2025-06-24-ai-as-the-new-study-partner-yes-we-can-help-students-learn-responsible-ai-use/ AI to the Rescue: https://www.chronicle.com/special-projects/the-different-voices-of-student-success/ai-to-the-rescue

The Chronicle’s article “AI to the Rescue” offers a bit more of, shall we say, a balanced view of how today’s college students are utilizing AI in and out of the classroom. It’s a balanced take on AI, noting that not as many students as some faculty fear are using AI merely as a shortcut to skirt academic rigor. Some students are learning to use it as their 24/7 personalized tutor, seeing it as an essential part of their scaffolding to bolster their learning, manage cognitive load, and help them navigate the challenging times we put them through in higher education. The narrative is grounded in student voices and survey data, painting a nuanced picture that challenges the prevailing trope of AI as merely a cheating engine. Thank you, Beth McMurtrie.

I particularly liked Abeldt’s story:

Abeldt began building her AI arsenal. She used ChatGPT and watched YouTube videos to break down complex topics. Perplexity, an AI-powered search engine, helped her find information. Google NotebookLM turned class notes and lecture slides into study guides and podcasts, which she listened to on the way to class. Quizlet made flashcards and ChatGPT created mock exams and quizzes. Abeldt even used AI to develop a study plan to avoid procrastinating or spending too much time on any one assignment. She followed its guidance, and it worked.

Abeldt’s outcome? A GPA rebound, and more importantly, a sense of agency restored.

Not all are on board, and nor should they be

Everyone — faculty, staff and students alike — should proceed with caution and not jump into AI without proper guidance and boundaries. As we know, OpenAI, Anthropic, Google, Meta, and the corporations involved in developing AI have zero interest in protecting humanity as they develop these tools. There is no doubt, it can cause harm, and already has. We’ll continue to see new examples of harmful things resulting from the irresponsible or naive use of AI increasing in frequency. Many should avoid it altogether. It’s not clear when it should be used, if at all. Despite its danger, we are noticing some positive outcomes for those who understand how to use it responsibly. Anecdotes are not data, but my colleagues and I have noted that in our higher-level courses and research with our students, AI can be quite helpful. However, even in these contexts, it needs to be used responsibly and with caution. (Responsible AI – I will keep pushing this with urgency, no matter what my colleagues think of me.) Some, particularly in the humanities or those with strong ethical stances, opt out entirely, citing concerns about bias, trustworthiness, or the environmental impact of large models. I respect those who take a stand through informed, data-driven decisions.

The article points out that some blame AI for the erosion of peer networks, as classmates default to AI tools for feedback and collaboration, bypassing the rewarding experiences (and challenges) that come through human interaction with their peers, meeting with TAs, and coming to the office hours of their instructors. No. Blame that on social media and mobile devices, which may both be considered bigger disruptions to regular human contact than anything AI is doing. AI might be making it easier to avoid human interaction, but social media will always surpass the damage done to promote human interaction than anything AI will ever do.

The fact is that very few campuses are monolithic; this is more true in the primarily undergraduate liberal arts colleges. The use of AI depends on the discipline. Let’s say that again – AI use and attitudes are shaped by discipline, workload, and personal philosophy. Let’s start respecting each other’s choices and autonomy to choose how they want to use it!

Here’s my take (subject to change as fast as AI is evolving):

Everyone should use AI only if they have a full understanding of how it works. There should be no judgment on individual choices made. Each user should understand how the model they are using is trained, with what data, and understand its strengths and weaknesses, as well as its ethical, societal, moral, and environmental impacts. If you determine AI is not right for you, your class, or your own learning, then take that stand, explain to your students, with data, why you are choosing that path to help them make future informed decisions, and make choices that are right for you. Let’s stop the virtue signaling, guilting, and condemning others who have chosen to use it. 

We all need to figure out what is right for our courses, our departments, and our programs. We need to teach our students to be ethical and responsible users of AI today. Every administration should be crafting a general policy on ethical and responsible AI use for their campus. But understand that there is no one-size-fits-all solution. That makes it challenging.

Disruption

The fact is, as I’ve stated elsewhere, AI is not going anywhere. To choose to ignore it is not a good way to move forward. It’s a colossal disruptor. There is no doubt that AI, like so many technologies over our lifetimes, is a disruptor that is displacing the norm of “how things are done.” It’s not the first, and it won’t be the last. Consider:

  • Industrial Automation, an endeavor around long before AI, has displaced traditional job markets, increased factory emissions, and disrupted landscapes from resource extraction.
  • Automobiles enabled great encomic growth, but led to urban spawl, traffic fatalities, and has been a formidable contributor to air pollution, also contributing to the need for substantial resource extraction for oil.
  • Nuclear Technology has such potential, but ask Chernobyl about the risks.
  • Computer and Internet Technologies have enabled nearly every modern disruption today! Want to point fingers at someone for AI? Well, blame this category! It disrupted how we communicate and do our work day-to-day. And, it caused incredible amounts of privacy and security concerns while continuing to contribute to the social and digital divides in the world. And the e-waste generated from new phones and tech being released year after year is profound. (It also contributed to enormous labor waste measured in countless hours of Solitaire played on the company’s dime!)
  • Biotech has had some remarkable improvements in crop yield at the cost of long-term health impacts, chemical runoff, loss of biodiversity, water pollution, and destruction of ecosystems.
  • Plastics, which we are only now seeing the huge negative impact on plants, animals and human health from microplastics. 

Disruption seems to be a historical norm. So, what is different about AI?

While many of the above are considered disruptors, AI is just… different. I’m not a philosopher by any stretch of the imagination. But, I think way too much, and I try my hardest to just listen to my colleagues across the college, combined with reading too much about AI. Here’s my take…

  1. Scope of human displacement – none of the above technologies have encroached on those humanistic things we do that are cognitive, creative, or fall under decision-making domains, the way AI has. It’s infringing on those things we thought were reserved for humans.
  2. Speed and Scale of change – the above disruptions took time to evolve. The pace that AI is moving is unprecedented, unpredictable, and driven mainly by large, greedy corporations who don’t give a %)@# how it impacts humans. And, it’s happening while we have an administration with no interest in curtailing and regulating how its used. This is downright unnerving and incredibly frustrating. As soon as I have something figured out, more capabilities emerge.
  3. Existential and Identity as Humans – AI is causing us to wake up in the morning, wondering what it means to be a human in the era of AI. 
  4. Trust, security, control, accountability, etc – AI systems are treated as those mythical “black boxes”, i.e., algorithms we blindly use with no concern about what’s going on behind the scenes. Who is responsible when AI causes harm, makes a biased decision, or generally has unintended consequences? 
  5. Privacy and manipulation – OK, so this really is part of the previous concern. But it bears repeating. AI can process an incredible amount of data, making it quite easy for “big brother” systems to monitor and perform regular surveillance without us knowing. Then there are the “deepfakes”, misinformation, etc… all posing greater risks than any of the above disruptors.
  6. Media – boy, is the mainstream media having a “hey day” with AI. Fear sells the news! News outlets love it when they can get your attention through fear and manipulation of your emotions. Facts? Pffft… not necessary when I got you emotionally charged. It’s what drives our political coverages on the news too. AI is being portrayed as a formidable, uncontrollable force, conjuring up images of roge terminators taking us out. The media fuels these anxieties.

There is a lot of fear and discontent over AI, rightfully so. It is indeed a disruptor, but unlike anything we have seen in quite some time, perhaps in human history. Or is it? Time will tell. For instance, automation was long criticized as a source of job displacement, but eventually realized it would create additional jobs. 

Balance – the key to everything

Let’s get back to the article at hand. Fortunately, it ends on a rather pragmatic note. The future of student success, as the author suggests, lies in balance – something we’re not very good at in academia, especially as we become a more polarized society. Can academics (me included) leverage AI for efficiency and accessibility, while doubling down on those things that make us human? Can we emphasize more intuition, compassion, innovation, creativity, and nuanced judgment? These are the things that AI cannot replicate. Can we reframe how we judge our students with our final grade assessments by rewarding their curiosity and willingness to challenge themselves and take intellectual deep dives instead of encouraging them to only learn enough to pass tests?  

Remember – AI can only imitate, not innovate. Let’s come together to figure out how to reimagine our mission in higher education – fostering the next generation of critical thinkers, world-changers, and ethical minds. It is essential if we are to maintain relevance.

Back to top ]]> AI GenAI https://www.eg.bucknell.edu/~brk009/posts/2025-06-24-ai-as-the-new-study-partner-yes-we-can-help-students-learn-responsible-ai-use/ Tue, 24 Jun 2025 04:00:00 GMT You mean, ChatGPT isn’t good for your brain? https://www.eg.bucknell.edu/~brk009/posts/2025-06-19-you-mean-chatgpt-isnt-good-for-your-brain/ https://www.theweek.in/news/health/2025/06/19/chat-gpt-might-not-be-good-for-your-brain-new-mit-study-finds.html

A recent MIT study raises concerns about how using ChatGPT might impact critical thinking and brain engagement. Generalizing the outcomes here (which have not been published or peer-reviewed yet):

  • Population size? n=54 – hardly enough to suggest much of anything on humanity, though it’s not the only study to suggest negative impacts on our brain activity from human overreliance on genAI chatbots.
  • An interesting aspect of the study was the use of an EEG to measure brain activity. Participants who used ChatGPT to write SAT essays showed the lowest brain engagement compared to those using Google Search or no tools at all. EEG readings revealed reduced neural activity. Not surprising at all.
  • Their essays were more uniform and lacked creative or original thought… as one would expect from any generative AI engine that serves as nothing but a fancy-ass stochastic regurgitation machine.
  • Another obvious outcome – participants also became [insert-shocked-face-here] lazier over time. When they didn’t use their own brain in their writing, they struggled to recall their own writing when asked to revise without the AI.

The group that didn’t use any tools demonstrated the highest neural connectivity, were more engaged and curious, and felt more satisfied with their work.

I have yet to talk with any academic in any field who is not sounding enormous alarms (not to mention frustration from cheating) from student use of AI. What I tell them is this: Generative AI, like it or not, is here to stay. Power-hungry, greedy corporations are driving it with no interest in understanding its impact on humanity. Big companies are lured by the temptation to save a buck or a million, and the tech companies are ready to serve the next AI plate. Like it or not, AI is not going anywhere. AI will continue to become more capable and yes, even awe-inspiring.

The cynical side of me (which is about 99% of me these days) views genAI as democratizing cheating. I’ve been privy to students cheating. All of us educators have. But, the best cheating came from the affluent – those students who could afford their subscriptions to cheating websites and services, some even paying for their code or essays to be written by cheating “farms” around the world! Well, ChatGPT has made cheating accessible for all! Snark aside, let’s be realistic -cheating has been around since the dawn of education. Stop worrying about the cheating! My God, that’s the least of our problems today. It doesn’t matter what fancy language you put on your syllabus, what rules the administration tries to put in place, or what types of lock downs you put on school computers – students are going to cheat, and they have access to all the best tools out there, and they likely know more than you do on how to use them.

My observation is that college students do not know how to use AI responsibly. They do not know how to utilize it to aid and promote their own learning in a way that keeps them just as engaged and invested in their own learning, while protecting their own critical thinking with the tools.

In theory, if learning with AI tools does not result in students yielding tangible, measurable evidence that their intellectual growth and critical thinking exceeds those measures without the tools, then what are we doing even having this discussion? Problem solved! But that’s not realistic. Humans (me included) are opportunistic, stressed-out creatures who will always have moments of needing to cut corners to get work done, and damn, generative AI makes that easy.

Our real problem – teaching our kids how to learn, how to protect their own brain development, in the era of highly-accessible generative AI tools.

It’s not going anywhere. It will be used. What do we do?

I am trying to believe that eventually the hype and chaos will subside in time, allowing it to become viewed as a tool, maybe even an essential tool. It will become much like the calculator for a mathematician, the stethoscope for a doctor, or the IDE for a software engineer. I hope that in time, genAI will mature to become an essential tool to help us with the mundane, repetitive tasks so that we can focus on getting more done, allowing us to become more efficient with our time so that we can leave each day accomplishing more than we could before, ideally letting us spend more time with our friends and families (you know, those things we used to do before mobile devices and social media. Remember having conversations at restaurants?) But, it’s going to take a lot of time to get there, and honestly, I’m not sure we (at least in the U.S.) have what it takes to enact the legislation urgently needed to regulate how AI is used in K-12, or even whether it should be allowed to be used at all in the early years. This is not a partisan issue. This is one of the most important pieces of legislation that needs to be passed to protect our future. GenAI in the hands of kids, at the age when their brains are still growing, is the absolute worst possible use of any technology today. We need to help our K-12 educators learn how to use the tech themselves and recognize early signs when kids may be using it. Every discipline at the college level needs to reimagine its curriculum to include instruction on how to teach students to use AI safely and responsibly while protecting their own critical thinking. Ideally, if done correctly, it should enable us to accomplish more in the classroom. Our students should be able to conquer more challenging problems than ever before… if and only if we help them learn how to use it! And that’s on us, friends – the educators of the world – to help our next generation students learn how to do that. Unfortunately, times are bizarre right now. With the recent devaluing of education in this country, I’m not sure we have what it takes to wake up and fund and support the education system in a way it needs to get through this time. Papers like the above will become the norm, to the point where it’s mostly AI that is publishing another paper announcing yet another test where it outperforms humans on yet another measure of brain activity, intelligence, critical thinking, etc.

Legislation isn’t going to happen, at least not in the near future. It’s on all of us, as educators and part of our society that values education, to lead the way in protecting our kids. In my courses, it’s my responsibility to teach my students how to utilize these tools in ways that promote their brain development, learn how to use AI responsibly and safely, and minimize harm to humanity. I don’t have answers. None of us does, but pretending like it’s not here and doing nothing is not the answer. Failure to do so will transform our society into some form of an idiocracy within the next generation, with humanity at the mercy of powerful AI engines that do all our thinking for us. (I never realized that Mike Judge was a modern prophet.)

I consider myself a huge fan of technology, dating back to my high school days. The technology itself does not scare me. I find it fascinating and am actively using AI and ML models in my own work. It has a purpose. What scares me is the willingness of society to let AI do their thinking for them. My worst fears are letting our kids use it with no reins, with no understanding of the harm its use is doing to their mental capacities in the long term. The damage is real, and if we don’t help our kids and college students learn how to use it safely, it will have devastating impacts on humanity.

This is a disruptive technology. Disruption causes mass chaos for a time as society adjusts. It also displaces and eliminates those who fail to evolve with it and adapt. Just ask Kodak. Or Blockbuster. Or how about Blackberry? Or Palm. MySpace anyone? AOL?

Back to top ]]> AI GenAI https://www.eg.bucknell.edu/~brk009/posts/2025-06-19-you-mean-chatgpt-isnt-good-for-your-brain/ Thu, 19 Jun 2025 04:00:00 GMT Workshop – Demystifying AI – May 23, 2025 https://www.eg.bucknell.edu/~brk009/posts/2025-05-23-demystifying-ai-workshop-may-23-2025/ The links below are to Python notebooks for each section of the workshop. Open the notebook, which will take you to a Google Drive link that is a read-only version of the notebook. Open the link, save a copy of the notebook file in your Drive space, and open it in Google Colab. You can also run the file locally on your own machine if you have a complete Python environment installed with Jupyter/JupyterLab or an editor that lets you edit Jupyter notebooks. All notebooks were tested on both Google Colab and natively.

These materials are part of workshops taught in my role as Faculty Fellow of the Dominguez Center for Data Science.

Workshop – Demystifying AI, Presented at Bucknell University, May 23, 2025. ©2025 by Brian R. King, licensed under CC BY-NC-SA 4.0. To view a copy of this license, visit https://creativecommons.org/licenses/by-nc-sa/4.0/

Back to top ]]> AI GenAI Workshop https://www.eg.bucknell.edu/~brk009/posts/2025-05-23-demystifying-ai-workshop-may-23-2025/ Fri, 23 May 2025 04:00:00 GMT Extended Reality (XR) https://www.eg.bucknell.edu/~brk009/posts/2024-11-03-extended-reality-xr/ I developed an intense interest in understanding how Extended Reality devices could be used for, well, “good.” Over the pandemic, as many of us were stuck home, I purchased a couple of the Oculus Quest headsets for my family. Frankly, we were blown away. Perhaps it was just the euphoria of being able to finally do something other than sit while watching the news about a virus taking over the world. We needed something to just help us get up, move around, and have fun. The technology was still considered early, but I was amazed at what it was capable of. (We would play the infamous Beat Saber, competing against each other for high scores! Even better – we were getting cardio exercise in while having fun!) Surely there must be a slew of use cases for this tech? So, I decided to dive in head first. I purchased the Quest 2 as Meta was taking over Oculus. Then the Quest Pro where we worked with colleagues at Geisinger Commonwealth School of Medicine to develop a couple of prototype VR applications where students could enter an OR simulation and learn basic cardiology while in remote areas. My students learned the ORamaVR platform to make developing their medical simulation prototype much easier than it would have been otherwise. I also started teaching in the EXCELerator program for the College of Engineering, where some incoming students come for an immersive experience in prepping for the college experience on campus over the summer before the start of their first semester. I work with a small group of students each summer helping them learn about design and development framework in engineering using VR. Lately the Meta Horizons creation platform has been quite good for that.

I’ve developed a brand new course (offered the first time in Spring 2024) where I teach students about Design and Development in XR using Meta Quest 3 headsets. Students learn about Unity for the first half of the semester by working through a wide range of custom labs and a suite of resources on the Unity Learn platform. The rest of the semester they spend their time going deeper as they implement various interactions and figure out how to create accessible, immersive experiences by working on their own projects. It’s been awesome watching students develop immersive worlds on their own from scratch! I’ve been amazed at what some students have been able to create. It’s a lot to cover in one semester, but the fun of working with these devices and developing in Unity has far outweighed the work, and I think most of the students have agreed.

If you’re interested, reach out to former students of the course, or just send me an e-mail and we can talk!

Back to top ]]> Teaching XR https://www.eg.bucknell.edu/~brk009/posts/2024-11-03-extended-reality-xr/ Sun, 03 Nov 2024 04:00:00 GMT Updating CSCI 205 https://www.eg.bucknell.edu/~brk009/posts/2019-11-23-updating-csci-205/ CSCI 205 has been a highly successful course for our majors. It is a lot of work for students, and likewise for the instructors who teach it (myself and Prof. Chris Dancy.) But, the rewards have been plenty, as the course teaches a lot about there is a lot of material that is out of date. The course will still relying on Java 7, and used Netbeans.

I’ve taken quite a bit of extra time this semester to update some of the course. As of Fall 2019, the course has been updated in the following ways:

  • We are now using IntelliJ IDEA
  • The course has been updated to Java 12
  • Many videos have been re-recorded to address the updated content, including:
    • Heavier emphasis on lambda expressions than ever before
    • Teaching more java.nio and java.nio2 along with java.io
    • Added new material on socket programming with java.net
    • Added new material on multithreading and concurrency in Java
    • Introduced the Java Stream API (not to be confused with the I/O streams)
  • The final project has been overhauled. Every student now must use Gitlab for all their Scrum task board and sprint management (This has worked surprisingly well!)
  • Expanded the JavaFX material 

It’s a start. There is a lot to be done still.

In a recent discussion with Chris Dancy, he expressed significant interest in incorporating elements of engineering social justice into the course. This represents a broader move by the engineering community at large to start helping our students recognize the impact that their choices have on humanity. I am at fault here. Like many of us, we focus on the goals without teaching our students the impacts that their choices have. Well, that’s not entirely true. I discuss impact – computational resource impact. That’s not enough. I do not give enough attention to social, moral or ethical impact. So, I believe this will be the next set of revisions we make to the course. Chris may likely start on some of those changes in Spring 2020. But, we’ll likely make a more substantial effort to incorporate this into the project over the summer. It’s time for engineers to emphasize people as more important considerations than profits.

Back to top ]]> CSCI 205 Java Teaching https://www.eg.bucknell.edu/~brk009/posts/2019-11-23-updating-csci-205/ Sat, 23 Nov 2019 05:00:00 GMT Past Research Projects https://www.eg.bucknell.edu/~brk009/posts/2019-07-13-past-research-projects/ The following are research projects that, for one reason or another, ended up falling down the priority list and are no longer being actively worked on. I list them here as a possible conversation starter with students looking for interesting work

  • [IN PREP] Cowen R, Mitchel MW, Hare-Harris A, King BR. Incorporation of Brown’s stages of syntactic and morphological development in a word prediction model of conversational speech from young children
  • [IN PREP] – Cowen R, Mitchel MW, Hare-Harris A, King BR. An adaptive n-gram based stochastic word prediction model for conversational speech.
  • [IN PREP]- Hare A, Essae E, King BR, Ledbetter DH, Martin CL. Determining the dosage effect of copy number variants in the human genome.
  • [IN PREP] – Ren C, King BR – Protein residue contact map prediction using bagged decision trees
Back to top ]]> Research https://www.eg.bucknell.edu/~brk009/posts/2019-07-13-past-research-projects/ Sat, 13 Jul 2019 04:00:00 GMT Current Student Research https://www.eg.bucknell.edu/~brk009/posts/2019-07-02-ongoing-student-research/ These are ongoing projects as of Summer 2019

Bhagawat Acharya ’20 – Using deep learning for handwriting text recognition.

  • This is a collaborative, interdisciplinary project with Katherine Faull (Comparative Humanities and German Studies) and Carrie Pirmann (Research Services Librarian). We are working together to develop an improved handwriting translation pipeline to increase the HTR throughput of 17-18th century Moravian handwritten literature that is part of the Moravian archives.
  • Funding – Bucknell Emerging Scholars Summer Research Program

Taehwan Kim ’20 – Using Deep Learning to Forecast Monthly Extreme Temperatures over the United States

  • Undoubtedly, climate change is one of most pressing, disconcerting issues of our time. Collaborating with atmospheric science and aerosol science expert Dabrina Dutcher, Assistant Prof. in Chemistry and Chemical Engineering, we are exploring the use of deep learning to develop advanced models that can improve future temperature predictions
  • Funding – Katherine Mabis McKenna Environmental Internship

Lily Romano ’20 – Software for Aerosol Analysis

  • We are developing a new software toolkit to aid in the aerosol research of my colleagues in Chemical Engineering, Dabrina Dutcher, PhD and Timothy Raymond, PhD. Lily is resuming work that was initiated by former student Khai Nguyen ’18 on the software, including advancing the data analysis tools available for aerosol researchers.
  • Funding – Clare Boothe Luce Research Scholars Program

Kartikeya Sharma ’20 – Trajectory Gaze Path Analysis on Eye Tracking Data for Autism Spectrum Disorder Studies

  • This is a collaborative project with my colleagues, Vanessa Troiani, PhD and Antoinette Sabatino DiCriscio, PhD at the Geisinger Autism Developmental Medicine Institute. The primary aim is to develop a toolkit for the eye tracking research community that incorporates my novel method for extracting scanpath trends from group-level eye tracking data.
  • Funding – Ciffolillo Healthcare Technology Inventors Program

Yili Wang ’21 – Using deep learning to identify discriminative features of images with high interest of autistic children

  • This is a collaborative project with my colleague Vanessa Troiani, PhD at Geisinger Autism and Developmental Medicine Institute. This is also a continuation of a project with former student Tongyu Yang `17, who is continuing to assist with the effort
  • Funding – Bucknell Program for Undergraduate Research (PUR)

These are projects that are unfinished for a variety of reasons:

Back to top ]]> Research autism deep learning eye tracking students https://www.eg.bucknell.edu/~brk009/posts/2019-07-02-ongoing-student-research/ Tue, 02 Jul 2019 04:00:00 GMT My joy with ILTM https://www.eg.bucknell.edu/~brk009/posts/2018-08-27-iltm-ending/ I had the joy of becoming a core faculty member of the Institute of Leadership in Technology and Management for the past two summers (Summer 2017, 2018). I found this to be one of the most transformative experiences available to Bucknell students since I’ve been here. I was honored to be part of this program. I worked with some absolutely wonderful students in ILTM! However, as a result of this opportunity, my scholarship was substantially halted for the last two summers. Thus, I have not taken on any new students for quite some time.

I was also on sabbatical during the entire 2017-18 academic year. During this time, I continued to work on interesting projects collaboratively with Dr. Vanessa Troiani at Geisinger Autism and Developmental Medicine Institute. As much as I’ve found much pleasure working in various areas of bioinformatics, I decided it was time for me to explore other areas of sequential data analysis. Dr. Troiani and her lab members have invigorated me with new opportunities in pattern mining mass quantities of eye-tracking data. This ultimately led to another collaborative project involving Dr. Troiani and our own Prof. Evan Peck. Slowly, the research agenda is ramping back up again. I applied to 5 different grant opportunities, of which, to date, one has been awarded, and a much larger one is currently under review.

I’ve also become more involved in interdisciplinary teaching and research opportunities across the university. Bucknell is at a point now where we can truly provide some very interesting transformative experiences to our students – rare opportunities that very few colleges can offer. To do so, however, we must leverage the opportunities that exist across disciplines. Thus, I’ve been intentional in my pursuits to identify new opportunities outside of my own department, and my own home – the College of Engineering. For instance, I’ve had great joy working with my collegue, Prof. Abby Flynt, on both teaching and research projects. (We both recently received the Presidential Award for Teaching Excellence for 2018, and co-mentored a wonderful student, Alexander Murph, who completed an honors thesis and is now at UNC Chapel Hill pursuing his PhD in Statistics!)

Speaking of new, unique opportunities for interdisciplinary work. I’m looking forward to seeing new things happen with our new College of Management, where I expect some interesting collaborations with new faculty who will be part of their new Analytics and Operations Management program. I’ve been spending time with them recently serving on a committee to help them hire new faculty for this exciting program.

Of course, I can’t forget our wonderful friends in Biology, who were so instrumental in collaborating on my bioinformatics projects very early on during my pre-tenure days here. Needless to say, there are great colleagues across this university, with lots of data! It’s a rich place for a data scientist!

Sequential data mining and analysis – it will always remain my primary area of focus, and it’s exciting to be able to afford the risks with tenure to be able to stretch my core interests toward new areas. Fortunately, sequential data are ubiquitous. Thus, I’ve branched away from biological sequence analysis and delved into numerous other areas of sequential data. I will update soon.

My post tenure feelings

So, is tenure all it’s cracked up to be? Well, I’m now in the midst of my third year post tenure. Or is it my second? I don’t even know. I’m burnt out, thanks to the vicious down side of tenure – SERVICE! Once a faculty member receives tenure, it seems as though you are put on a list by the administration throughout the college and university. This list is special. I believe the title of the list is, “PEOPLE WHO WE CAN GUILT INTO SERVING ON COMMITTEES NOW THAT THEY HAVE TENURE.” This semester, I honestly have lost count of the committees and the other opportunities where I have said “yes” to volunteer for opportunities to help my colleagues. The result, I regularly have a minimum of 10-15 additional hours per week dedicated to service obligations; that has recently reached 20+. Those are hours that are on top of my normal teaching hours in and out of the classroom, which are easily 40+ hours, and that doesn’t include my normal teaching/service duties, such as academic advising, department meetings, mandatory caffeine pursuits, and so on. (An academic has no concept of a 40-hour work week. It doesn’t exist.) This is a huge challenge that I’m struggling with. It is due, in part, to a very young, vibrant department of faculty who are going through the tenure process. Thus, the relatively few of us who have tenure take on a lot of the service obligations to protect them as they work through tenure. And, of course, I know, I know… the real reason? Because I often find it difficult to say, “NO!” Like I said, I work at a great place with wonderful colleagues. I believe it’s important to pay it forward. I had people senior to me who were once in my shoes and protected me from excessive service obligations, and I will do the same. The challenge is the imbalance in the department. We’ve had a lot of people retiring in recent years. In time, the balance should be back to normal as others get through the tenure review process, and they can share the service burdens.

Anyway, the most important thing that has me excited? First, I’m teaching BOTH a data mining AND a data science course this Spring! Second – this summer 2019 is mine! All mine! [Insert-evil-laugh-here]. I have not had a summer for research since 2016. So, I have several projects that are ramping up, and I am looking for new students to work with me this summer. Funding is available. Send me an e-mail if you are interested.

Back to top ]]> Bucknell ILTM interdisciplinarity https://www.eg.bucknell.edu/~brk009/posts/2018-08-27-iltm-ending/ Mon, 27 Aug 2018 04:00:00 GMT