AI Transparency Deadline Approaching

Starting August 2, several provisions of the EU AI Act become enforceable, including requirements for transparency when AI systems interact with people and for AI-generated content to be marked as artificial. An exception exists for AI systems used in law enforcement, with safeguards for fundamental rights. Certain AI systems already on the market before August 2 will have a grace period until December 2 to comply with watermarking obligations.

• Transparency is required for AI systems interacting directly with natural persons, unless obvious or used for law enforcement.
• AI-generated audio, image, video, or text content must be marked in a machine-readable format, with exceptions for standard editing or law enforcement use.
• Deployers of emotion recognition or biometric categorization systems must inform those exposed, unless used for law enforcement.
• Deepfake content generated or manipulated by AI systems requires disclosure, with exceptions for artistic works or law enforcement.
• Some AI Act obligations, particularly for high-risk AI systems, have been postponed due to the Digital Omnibus on AI approval.
  Continue reading https://www.luizasnewsletter.com/p/ai-transparency-deadline-approaching