EU AI Act adopts transparency code, bans AI‑generated non‑consensual intimate content

The European Union has introduced several new measures to tighten the AI Act. On 24 July 2026 the EU published Regulation (EU) 2026/1744, amending the AI Act to prohibit AI‑generated non‑consensual intimate images, including nudification tools, and to require labeling of all synthetic content by 2 December 2026. The same week the European Commission released guidelines that clarify transparency obligations for AI providers and deployers, effective 2 August 2026. The rules mandate machine‑readable marks on AI‑generated or manipulated content, user notices for interactive systems such as chat‑bots, and clear labeling of deepfakes and AI‑generated material on matters of public interest.

Google announced on 24 July 2026 that it has signed the EU’s Code of Practice on Transparency of AI‑Generated Content and is extending its SynthID watermarking technology across a consortium that includes Apple, OpenAI, ElevenLabs, Kakao and NVIDIA. SynthID embeds invisible signals into AI‑generated text, images, audio and video, allowing downstream tools to detect the origin of the content while surviving typical edits. The code provides a voluntary framework for compliance, and signatories will have a narrow window to align their systems before the August deadline.

Together, the amendment, the new transparency guidelines and industry commitments aim to improve accountability, protect individual dignity, and give users clearer information about AI‑generated media across Europe.