AI’s Literary Blind Spot: GPT Models Praise Nonsense, Study Finds
New research reveals OpenAI’s GPT models rate nonsensical text as high-quality literature, raising concerns about AI’s reliability in aesthetic judgment and decision-making processes.
AI Evaluation
Meta Unveils AI That Evaluates Its Own Work, Paving Way for Autonomous Agents
Meta releases new AI models, including a Self-Taught Evaluator that uses AI to evaluate AI, reducing human involvement and paving the way for autonomous AI agents capable of self-improvement.