Meta Launches AI That Can Monitor Other AI As Human Involvement Diminishes

Meta has released new AI models, including a “Self-Taught Evaluator” that could reduce human involvement in AI development. This model uses AI-generated data for training and employs a “chain of thought” technique to improve response accuracy. The technology potentially paves the way for autonomous AI agents capable of self-improvement. Meta researchers suggest this approach could eventually surpass human evaluation in effectiveness. The company also released other AI tools, including updates to image identification models and datasets for material discovery. Unlike some competitors, Meta makes these models publicly available.READ FULL