Meta’s AI Detector Can’t Detect Images It Generated Itself, Report Finds
AI Today Summary
Meta introduced an AI image generation model called Muse Image along with a watermarking system named Content Seal, designed to identify AI-generated images even after modifications like cropping. However, a Reuters report found that Meta's AI detection tool failed to recognize over half of the cropped images as AI-generated, highlighting ongoing challenges in AI content detection. This shortfall underscores the broader difficulty in keeping pace with the rapid growth of AI-generated deepfakes,,
Covered by 1 source
AI Today publishes original summaries and links out — never the full source article.