AI Frontiers reports that new transparency regulations in the European Union and California, which took effect on August 2, 2026, now require generative AI systems to automatically attach machine-readable metadata — known as Content Credentials or C2PA — to synthetic audio, images, video and text. The rules aim to help distinguish authentic material from AI-generated content and could eventually assist courts in evaluating deepfake evidence. The piece notes that starting in 2028, camera manufacturers will also be required to build in provenance records by default, establishing a broader global baseline for content transparency.
