Meta announced the launch of a new AI model, called 'Self-Taught Evaluator', which reduces the need for human involvement in AI development.

The new AI model, called the 'Self-Taught Evaluator', uses a technique called 'chain-thinking', similar to that used by OpenAI's recently released o1 models, to make reliable judgments about the responses of AI models. This evaluator was trained exclusively with AI-generated data, completely eliminating human intervention.