Because there is no general form of “AI,” but rather sets of models, contexts, and tools, there is no one debate or conversation about (choose your adjective) “trustworthy / responsible / safe / ethical” AI.
The report of the Committee on Trust in Higher Education reflects AI’s Janus-faced presence in student learning.
On the one hand, they believe “Yale must recognize the epochal advances in technology and artificial intelligence, and ensure that graduates are prepared to deploy, design, and improve these tools.”
On the other, that “Whatever its promise, AI in its current use on campus undermines the expectations of focused, disciplined thinking that have long been the standard features of a rigorous education.”
It follows that learning to design and deploy AI tools must itself result from disciplined thinking. This page offers conversations and prompts to promote that thinking. It foregrounds existing Yale expertise as a resource for your own explorations.
How would you feel if your instructors started automating certain tasks that they are responsible for?
How does that compare to the quality of work that you generate without any AI assistance?
Discussion Prompt: Is there a way to use AI without being “seduced”? Does college culture need to change for AI not to be so seductive?
Writer Meghan O’Rourke (Yale English) observes the tension between AI’s personal utility and pedagogical harm: students who lean on it will skip the “meandering,” effortful process by which writers actually think on the page. AI’s “seductive cocktail of affirmation, perceptiveness, solicitousness and duplicity” and capacity for superhuman labor accelerates the high-achievement culture of college.
Discussion Prompt: Would developing AI philosophers be “democratization” of wisdom or a misunderstanding of the nature of philosophy? In your experience, can AI help you reason through complex moral, ethical, or logical questions?
Faculty in Yale’s Philosophy department recently published counterarguments about whether the world could benefit from AI’s philosophical reasoning abilities (i.e. can everyone now have a pocket philosopher).
Reflective Prompt: Is there a topic you’ve studied primarily through AI? Can you tell how your understanding of that topic differs from one you’ve spent substantial time exploring without AI?
Discussion Prompt: Could AI help scientists reach a broader audience, or will it weaken their rhetorical range and power?
Lisa Messeri (Yale Anthropology) and M.J. Crockett (Princeton Psychology) have argued that AI leads to “Illusions of understanding” and monolithic thinking among disciplines of scientific research.
In another published forum they debate the value of AI for scientific communications, they understand the hope that AI can “translate research findings for different audiences and ease the burden on already overtasked researchers” but worry that “Tasking generative AI with science communication threatens to erase diverse interpretations of scientific work.”
Reflective Prompt: How have you noticed the usage of AI affecting your learning experience? Can you imagine a study strategy that will ensure that 1) sufficient time is spent, and 2) AI use actively increases cognitive challenge? How do you know when you’ve experienced meaningful cognitive challenge?
A groundbreaking study of Chinese secondary education shows a sharp reduction in homework time and equally negative impact on test scores among student AI users. Only a tiny percentage of students appear to have benefited by “using the technology intelligently.” Notably, these students spent as much or more time on homework as non-AI users.
Discussion Prompt: Would you like to know whenever AI has been used? Can you tell already? How has the prominence of AI text (sometimes pejoratively called “slop”) affected how you write and read?
EU regulation is leading AI companies to develop durable “watermarks” for AI generated media.
This page will help you connect with colleagues thinking at the frontier of teaching with and researching AI.