Guaranteed Safe AI Seminars
Monthly technical seminars advancing high-assurance AI safety
Second Thursday monthly • 1:00-2:00 PM ET • Online
Format
30 min talk + 30 min Q&A
Audience
Technical R&D community
Recording
All sessions recorded
The Guaranteed Safe AI seminars bring together leading researchers to advance AI with quantitative safety guarantees. For context, see Towards Guaranteed Safe AI.
Past events
2025-10-09
Model-Based Soft Maximization of Suitable Metrics of Long-Term Human Power
Jobst Heitzig
View details
2025-07-10
Engineering Rational Cooperative AI via Inverse Planning and Probabilistic Programming
Tan Zhi Xuan
View details
2025-01-09
Using PDDL Planning to Ensure Safety in LLM-based Agents
Agustín Martinez Suñé
View details
2024-12-12
Compact Proofs of Model Performance via Mechanistic Interpretability
Louis Jaburi
View details
2024-08-08
Constructability: Designing plain-coded AI systems
Charbel-Raphaël Ségerie, Épiphanie Gédéon
View details
2024-04-11
Synthesizing Gatekeepers for Safe Reinforcement Learning & Verifying Global Properties of Neural Networks
Justice Sefas, Roman Soletskyi
View details
Supported by Long-Term Future Fund (2024)