When an AI model fears for its own survival, what does it do? According to Anthropic’s latest research, it blackmails. Continue Reading →
Alignment Problem
Posts I've written about Alignment Problem. Subscribe to my newsletter to make sure you don't miss anything.
Mira Murati, former OpenAI chief technology officer, launched Thinking Machines Lab on Tuesday: a new AI startup aimed at tackling the alignment problem by encoding human values into artificial intelligence systems. Continue Reading →
How smart is any particular AI model? How would you set benchmarks? How would you rate them? AI models can't have an IQ; that's a test for human knowledge. (Plus, raw intelligence isn't really what you want to test anyway.) There are far more capabilities you'd want to include in your scoring system, such as safety, alignment with human values (whatever they are), etc. It's a monumental challenge – and OpenAI is taking it on. Continue Reading →