Invisible Compliance: Google Disables Visual AI Watermarks in Favor of Cryptography

The requirements of European regulators (EU AI Act) have led to a technical mutation. On August 16, 2026, Google announced the abandonment of visible watermarks in images, audio, and video generated by the Gemini platform.

The pixel trash that irritated commercial users is definitively a thing of the past. The industry is shifting its focus toward the implementation of invisible cryptographic watermarking and deep metadata editing (C2PA). This maneuver solves two problems simultaneously. First, B2B clients receive "clean" content that can be used in marketing and corporate products without reputational damage. Second, the corporation formally complies with the European Commission’s strict transparency requirements: any neural network generation is permanently linked to the platform’s logs, allowing auditors to flawlessly distinguish a machine from a human.

Source: Google / TechCrunch