Kritia Law WatchAI-generated
EU AI Act transparency, implementation update and an Extia fine
AI-generated. This briefing was drafted by artificial intelligence and passed Kritia’s automated source and claims checks. It did not receive human editorial review before publication.
The European Commission reports AI Act transparency enforcement and an AI Board implementation update. The EDPB/CNIL reports a EUR 300,000 CNIL fine against Extia over erasure requests.
01 — ENFORCEMENT
Decisions authorities and courts actually published.
Failure to respect the rights of individuals: the CNIL fined EXTIA 300 000 EUR (2 minute read)
European Data Protection Board / CNIL· EU
The CNIL’s restricted committee imposed a EUR 300,000 administrative fine on Extia, according to the European Data Protection Board / CNIL record. The record says that, of 265 erasure requests received in 2024, more than three quarters had not been dealt with or had not been dealt with satisfactorily; 166 people were not informed of the action taken and 27 were informed late.
Why it matters to you
The record concerns a French company’s handling of erasure requests from candidates and former employees. It identifies failures involving processing requests and informing individuals of action taken.
02 — RULES & GUIDANCE
New or changed official text, guidance and proposals.
European Commission initiates enforcement of Artificial Intelligence Act rules and transparency standards (2 minute read)
European Commission — Interoperable Europe· EU
The European Commission — Interoperable Europe article reports that the Commission’s AI Office and national competent authorities started enforcing the Artificial Intelligence Act on 2 August 2026, including mandatory transparency requirements for interactive technologies. It says chatbots must notify individuals when they interact with AI, while deepfakes must be labelled and synthetic or altered content must carry machine-readable marks.
Why it matters to you
For firms using AI products, the article identifies transparency features reported as being enforced for interactive technologies and AI-generated or altered content. The article also says tools and reporting channels are available for possible violations or breaches.
03 — PRACTICAL
Worth doing, or worth the time to read.
AI Board holds its ninth meeting (2 minute read)
European Commission· EU
The European Commission’s news article reports that the AI Board met on 17 September 2026 under the Irish presidency of the EU Council to discuss EU and international AI policy, enforcement and implementation. The Commission presented an update on recommendations to support AI literacy and measures supporting transparency rules applicable from 2 August, including the code of practice and guidelines.
Why it matters to you
The article signals ongoing EU-level implementation work relevant to businesses using AI products. It describes recommendations and measures as under development or support activity, not as a new business-specific obligation.
04 — THE WIDER VIEW
Written about the rules, not by the people who make them.
Nothing worth your time today.
05 — COUNTDOWN
74days · 2026-12-02
Machine-readable marking for generative systems placed on the market before 2 Aug 2026.
What this isn’t
News about the law, not legal advice, and not an assessment of your business. Items marked Analysis are someone’s view of the rules, not the rules. Whether a duty applies to you depends on what you actually do. The public guides explain the general rules and link their primary sources.
Get this each publishing day
Free by email, with no account and no card. About a minute to read, every reported item linked to its source.