Bonn: The AI Act fines procedure now has a draft rulebook, and companies have until 9 November 2026 to comment on it. On 12 October the European Commission opened a four-week public consultation on a draft implementing regulation, reference Ares(2026)9589403, that sets how the Commission runs proceedings under Articles 75c and 75d of the AI Act. The Digital Omnibus on AI inserted those articles.
According to ppc.land, which reviewed the draft, the text has nine articles in four chapters. It sets a five-year limitation period for imposing fines and for enforcing them. Six kinds of interrupting action restart the clock, and a ten-year absolute cap applies. Adoption is planned for the fourth quarter of 2026, and the rules would take effect 20 days after publication in the Official Journal.
The draft also covers the rights of companies under investigation. Access to the Commission’s file comes only after preliminary findings under Article 2. Businesses that object to the disclosure of trade secrets get one-week windows to respond under Article 3. Advisers who work on a case face a three-year restriction on joining the company concerned or its competitors.
The stakes of AI Act fines come from the size of the penalties. Under Article 99 of the AI Act, breaches of the prohibited practices can draw fines of up to €35 million or 7 percent of worldwide turnover, and most other infringements carry up to €15 million or 3 percent. The Commission’s power to fine providers of general-purpose models has applied since 2 August 2026.
The Omnibus also softened the rules on staff training. Under Article 4, providers and deployers must take measures on AI literacy but no longer have to guarantee a set literacy level. A new Article 4a gives a legal basis to process sensitive personal data for bias detection and correction. Both changes affect how companies prepare for AI Act fines once inspections begin.
Enforcement of AI Act fines has begun in practice. Reed Smith reported on 26 October that the AI Office has sent its first requests for information to leading AI providers. The note names no companies and gives no reply deadline, and sources disagree on whether the requests went out on 14 October or earlier, so the timing remains unreconciled.
The Digital Omnibus on AI also reset the compliance calendar. It entered into force on 27 July 2026 as Regulation (EU) 2026/1744. Two new prohibitions apply from 2 December 2026, and the Annex III high-risk requirements move to 2 December 2027. The consultation will show whether providers think the draft protects their defence rights, and the final text will define how the AI Act fines regime works in practice.
