Reading groups
Read AI safety papers and compare notes on their arguments, evidence, and limits.
A community in Ann Arbor working toward
safe, beneficial AI.
How can we make powerful AI systems safe and beneficial?
A4 brings people in Ann Arbor together to study that question. You don’t need a technical background to take part.
Organized by Alan Tan, Luna, Claire Mao, and Aidan Lim.
Start with a system that follows a rule but misses the point. Then compare other failures and examine what researchers can learn by looking inside a language model.
Follow the reading pathRead AI safety papers and compare notes on their arguments, evidence, and limits.
Bring a paper or a question you want help thinking through.
Try out an idea and document what you learn about making AI safer.
Supported by