News
AI assessors says current science hasn't caught up to the safety measures people want
Anthropic and OpenAI say they are working with third-party evaluators to make sure their products are safe. Some evaluators caution that current AI science lags behind what people are calling for.