Technical AI Safety Fellowship

AISA is excited to announce the next iteration of our Technical AI Safety Fellowship—an introductory program covering the most important questions at the heart of AI safety and alignment research.

Through 90-minute weekly discussions over 8 weeks, fellows will have the opportunity to explore technical challenges in understanding, supervising, and shaping the goals of increasingly powerful AI systems, covering key topics including:

• Reinforcement learning from human feedback (RLHF)

• Reward misspecification and goal misgenrealizatiom

• Mechanistic interpretability and model evaluations

• Unlearning, red-teaming, and scalable oversight

See here for our Fall 2025 curriculum.

Applications for Spring 2026 are now open, and will be reviewed on a rolling basis!
The deadline to apply by is January 30th, 2026 by 11:59pm EST.