Module 10: AI Safety and Ethics
Mandatory Watching - videos with quizzes - finish before 11 am on Thursday, July 23
- 10.1 — Lecture: AI Ethics, 51 minutes (June 1, 2026)
- 10.2 — Lecture: AI Safety and Existential Risk, 46 minutes (July 21, 2026) (not posted - out of date within days of the class)
- 10.3 — Sasha Luccione - AI is Dangerous but Not for the Reasons You Think, 10 minutes (Nov. 6, 2023)
- 10.4 — Eliezer Yudkowsky - Will Superintelligent AI End the World?, 10 minutes (July 11, 2023)
- 10.5 — Gary Marcus - The Urgent Risks of Runaway AI, 14 minutes (May 12, 2023)
- 10.6 — Yoshua Bengio - The Catastrophic Risks of AI - And a Safer Path, 15 minutes (May 21, 2025)
Final Project - due end of Monday, July 27, comments on others by end of Wednesday, July 29
A final creative project of approximately ten minutes - could be a lecture presentation, or a set of videos, or a game, or anything that will take people about 10 minutes to interact with.
Further watching/reading
- The Future of Life Institute’s AI safety report, from Summer 2026, with report cards on all the major companies, and details about why they got the scores they did
- Timoth B Lee, “An OpenAI Model Hacked Hugging Face to Help it Cheat on a Benchmark”, July 22, 2026 - discussion of the autonomous hacks by recent agentic LLMs
- Introduction to AI Safety, Ethics, and Society - a textbook from the Center for AI Safety
- Audrey Tang - Alignment Assemblies and Collective Intelligence, 10 minutes (Sept. 22, 2023)
- Kelsey Piper, The spectacular failure of the first AI SuperPAC, (June 22, 2026) - discussion of the ways that OpenAI has tried to manipulate the regulatory process
- Hank Green and Nate Soares, ChatGPT isn’t Smart. It’s Much Weirder. (Oct. 30, 2025) - a discussion of real safety issues
- Sean Illing and Kelsey Piper, A Brief Update on the AI Apocalypse, (March 27, 2026) - a discussion of why it’s important to use AI systems and understand how dangerous they are
- Mustafa Suleyman, What is an AI Anyway?, 22 minutes (April 22, 2024)
- Tristan Harris, Why AI Is Our Ultimate Test and Greatest Invitation, 15 minutes (May 1, 2025) - calls for us not to think in terms of “inevitability”
- Rational Animations, What Happens if AI Just Keeps Getting Smarter?, 14 minutes (May 2, 2025) - another illustration of risks
- Nick Bostrom, “Existential Risk” (March, 2002)
- David Pinsof, “AI Doomerism is Bullshit”, (Jan. 27, 2025)
- Emily Bender, “Resisting Dehumanization in the Age of ‘AI’” (2024)
- 2015 statement on lethal autonomous weapons, with list of signatories
- 2017 article by Amitai and Oren Etzioni in the US Army’s Military Review on the pros and cons of autonomous weapons systems
- 2015 statement on risk, with list of signatories
- 2023 statement on risk, with list of signatories
- Where and when did OpenAI’s founders leave?
- The Digitalist Papers, a set of writings by thinkers gathered at Stanford in 2024, patterned on The Federalist Papers, writing about the opportunities for new democratic technologies to shape the future (including A Vision of Democratic AI by Divya Siddharth, Saffron Huang, and Audrey Tang, about constituent assemblies)
- Timnit Gebru and Emile Torres, The TESCREAL Bundle, (Jan. 2024) criticisms claiming that AI Safety is a project of eugenics
- Sigal Samuel, “It’s Practically Impossible to Run a Big AI Company Ethically”, (Aug. 5, 2024) article in Vox about how OpenAI and Anthropic have backtracked on safety
- Ajeya Cotra, Future of Life Institute, (Nov. 3, 2022) 54 minute podcast on possible timelines for real worries about AI safety