A curated collection of resources on AI in teaching, research, and practice, assembled for the Questrom faculty workshop. Organized into three tracks and ordered by how often we find ourselves returning to each one.
Assignment design, assessment, cheating, and what actually happens when AI walks into the room. Start with Mollick; the rest fill in the debate.
The default reference for faculty experimenting with AI in classrooms. Mollick's prompt libraries and "co-intelligence" framing now show up in real syllabi.
Ethan and Lilach Mollick's published syllabus and prompt library, actively used by instructors as a starting point for their own materials.
The Wharton working paper reporting that unrestricted GPT-4 tutoring boosted practice performance but hurt later learning, while guardrails reduced the downside. The single most-cited negative-effects study in the faculty discourse.
An Insights@Questrom piece by Chris Dellarocas, BU Associate Provost for Digital Learning & Innovation, arguing that generative AI disrupts traditional assessment before educators can redesign it.
HBS is integrating AI simulations, avatars, and live exercises into the case method, and giving MBA students access to ChatGPT, Harvard AI Sandbox, Claude, Claude Code, Lovable, Julius AI, Manus, and Gamma.
Google's announcement of NotebookLM features built specifically for students: study guides and learning tools grounded in a student's own sources.
Deming argues that AI's personalization benefits are real, but learning still requires guardrails and self-restraint from the student.
A preprint reporting that a designed AI tutor beat an active-learning classroom on short-run learning gains and engagement.
Report on how AI use in education can offload cognition, but can also be structured as tutoring, Socratic questioning, and verification.
An AI case-prep platform that runs adaptive voice-to-voice conversations with students on class cases and discussion questions before class, piloted in the core Strategy Formulation course for first-year MBA and EMBA students.
Cope, Frankenreiter, Hirst, Posner, Schwarcz, and Thorley test whether LLMs can grade law-school exams reliably enough to assist or partly replace professors.
A report on professors adopting oral exams as a response to AI-enabled cheating.
Panos Ipeirotis describes how to use an ElevenLabs voice agent and an LLM grading council to run scalable oral exams.
A professor argues that AI has changed classroom practice in some constructive ways, not only destructive ones.
Business schools are using AI-driven cases and simulations to teach management, reshaping how the case method is delivered.
An argument that colleges need to redesign assessment around work that is harder to outsource to AI.
An argument that AI raises the value of humanistic judgment even as it scrambles humanities teaching and research.
Instructors reviving low-tech assignments, typewriters included, to limit AI-assisted cheating.
Colleges are bringing back handwritten blue books to curb AI-assisted cheating.
Stanford MBA students who feel AI is outpacing the curriculum and changing the value of the degree.
Victor Kumar on why ChatGPT should scaffold thought rather than replace reasoning altogether.
Kumar proposes course redesigns that preserve student writing while reducing the incentive to hand it off to ChatGPT.
Gary Marcus argues that the costs of mass AI cheating are being dumped on educators rather than the companies enabling it.
A teacher's view of homework cheating, anti-AI backlash, and why classroom use of chatbots feels so destabilizing.
On professors using ChatGPT themselves, and the unease or backlash that can follow.
A Bluesky post excerpting professors' stories about students normalizing ChatGPT use and objecting when assignments are made AI-resistant.
A Bluesky pointer flagging an article as especially worth reading for people following the AI-in-education discussion.
A poem reflecting on what is lost when a student lets AI write the paper for them.
Concrete ways researchers are pulling AI into empirical work, literature review, and writing. Heavy on economics and Claude Code because that's where the toolkits are most mature right now.
A public CLAUDE.md showing how Andy Hall configures Claude Code for a research workflow.
A Marginal Revolution post excerpting Hall's argument that AI tools now belong in economists' empirical research toolkit.
A short MIT note laying out Andrews's advice on AI, skill-building, verification, and strategy for research careers.
A companion site to Korinek's NBER working paper that walks economists through building AI agents for literature review, econometric coding, data fetching, and multi-step research workflows, with hands-on LangGraph examples.
Karpathy recommends that researchers build topic-specific LLM knowledge bases that function like personal wikis.
A curated knowledge base organizing articles, concepts, and visualizations on using AI in business and economic research.
A catalog of reusable AI skills for economists covering literature reviews, data work, econometrics, writing, and engineering tasks.
Chris Blattman's practical guide for non-coders, sharing Claude Code workflows, templates, and skills for professional knowledge work.
An open-source Claude Code scaffold for empirical economics research, from literature review through journal submission.
A long-form guide to splitting academic work between Claude Code (for building) and Cowork (for organizing, formatting, and distributing).
A GitHub markdown prompt for having Claude Code build, compile, and repeatedly review polished Beamer slide decks.
Anthropic on how physicists use Claude for exploratory, conversational scientific work they call "vibe physics".
An X thread claiming Claude Code can replicate and extend an empirical political science paper with a mostly automated workflow.
A mathematician describes using ChatGPT to help produce a mathematical proof, step by step.
An AI-powered academic search engine that finds, summarizes, and cites results from a large research-paper corpus.
An AI medical-evidence platform that helps clinicians search, synthesize, and apply research and guideline information.
Field notes from outside higher ed: how AI is landing in work, code, culture, and everyday life. Useful context for what students are walking into.
The most reliable running log of what LLMs can actually do week to week, used heavily by people building real workflows.
NYT Daily on how AI coding tools are reshaping software engineering jobs and workflows.
An Insights@Questrom essay arguing that as AI automates routine coding and analysis, the real differentiator for data professionals becomes disciplined problem-solving: defining problems clearly, planning methodically, and guiding AI tools toward meaningful solutions rather than accepting outputs blindly.
A WSJ beginner's guide to choosing a chatbot and writing better prompts for useful results.
A Google Cloud roundup of real-world generative-AI use cases across industries and business functions.
A slide deck introducing Claude Code skills, how they work, and where they fit in an agentic workflow.
A Claude skill for generating polished slide decks with modern frontend web tooling instead of PowerPoint templates.
An InstantDB essay explaining how its real-time backend is built to support AI-coded full-stack apps.
A Reddit discussion about teams acting as managers of multiple coding agents rather than hand-writing most code.
A thread on the shift from directly writing code to supervising AI agents that do more of it.
An X video showing a tradesperson using AI tools to build niche software for his own workflow.
Younger workers are reconsidering careers as AI reshapes employer expectations.
Managers are pressuring employees to adopt AI tools or risk falling behind at work.
A roundup of reader-submitted examples of how people are actually using AI in everyday work.
A New Yorker feature on companion-AI products and the kinds of relationships people are forming with them.
A resurfaced calculator-era panic, useful as a historical analogy for today's AI backlash.