EU AI Act passed + LinkedIn integrates C2PA

The EU AI Act gets final green light by the Council of Ministers and requires labelling of AI content by May 25

The EU AI Act, the world's first comprehensive AI regulation, has been approved and will come into effect within this month.

This legislation introduces stringent regulatory requirements for AI systems, focusing on risk-based classification, data governance, and compliance.

AI generated content is categorised as “limited risk” and labelling and disclosure measures will become obligatory within 12 months.

AI systems (ChatGPT, Midjourney) and distributors (media, social media) will need to comply in regard to AI generated & enhanced content:

• Disclosure: AI-generated or manipulated content, such as deepfakes, must be clearly identified. This means all content that has been created or altered using AI technologies must be disclosed.

• Traceability: There will be an emphasis on traceability, requiring that AI systems used for content creation or manipulation maintain logs and records of their activities. This ensures that any AI-generated content can be traced back to its source, improving accountability.

• Standardization: The Act promotes the development of machine readable and common standards for AI content.

• Informed Consent: Users must be informed when they are interacting with AI systems and when content they are consuming is AI-generated or AI-enhanced.

CNBC, Reuters, Official EU website

LinkedIn adopts C2PA Standard for Content Attribution (AI Labels) following TikTok

LinkedIn has partnered with the Coalition for Content Provenance and Authenticity (C2PA) to label AI-generated content on its platform.

This initiative will add a new icon to any media with C2PA metadata, helping users trace the origin and history of such content, thus enhancing transparency and authenticity​.

This follows TikTok with a similar implementation a few weeks ago, further cementing C2PA as the standard for content attribution. Link

Scarlett Johansson and OpenAI fight over likeness of ChatGPT 4o’s voice “Sky”

Scarlett Johansson has raised concerns over an OpenAI ChatGPT voice named "Sky," which she claims sounds very similar to her own.

Johansson stated that she had declined an offer from OpenAI to use her voice for the project but was shocked to hear a similar voice in the final product.

Following her complaints and legal action, OpenAI agreed to halt the use of the Sky voice, clarifying that the voice was not intended to mimic Johansson and was from a different actress​ that was signed weeks ahead of Johansson’s rejection. Link