Schedule
Spring 2027 · Mondays, 3:00–6:00 PM
Readings will be posted later. The schedule is subject to change; updates are announced on Canvas.
Weekly schedule
| Week | Date | Topic | Due |
|---|---|---|---|
| Week 1 | Mon 1/11 | What Is AI Safety? How Large Language Models Are Trained | HW1 (TBD Nov 11, 2026) |
| Mon 1/18 | Martin Luther King Jr. Day, no class | ||
| Week 2 | Mon 1/25 | Reward, Alignment & Human Preferences | HW2 |
| Week 3 | Mon 2/1 | Data, Bias & Evaluation Pitfalls | |
| Week 4 | Mon 2/8 | AI Safety in the Real World: Case Studies | HW3 |
| Week 5 | Mon 2/15 | Red-Teaming & Jailbreaks | |
| Week 6 | Mon 2/22 | Interpretability | One-page project proposal |
| Week 7 | Mon 3/1 | Formal Verification | |
| Week 8 | Mon 3/8 | Scalable Oversight | |
| Mon 3/15 | Spring Break, no class | ||
| Week 9 | Mon 3/22 | Evaluations at Scale | Abstract submission |
| Thu 3/25 | Project submission & peer review start | ||
| Week 10 | Mon 3/29 | Guest Lecture (TBD) | |
| Week 11 | Mon 4/5 | Governance & Policy | Peer reviews due & discussion start |
| Week 12 | Mon 4/12 | Economic Incentives & Frontier Risks | Discussion ends |
| Week 13 | Mon 4/19 | Model Psychology, Constitution, Persona & Welfare | Camera-ready project, final poster & slides |
| Wed 4/21 | 🏆 Oral presentation announcements | ||
| Week 14 | Mon 4/26 | UT AI Safety Symposium (attendance required) | |