Digital Omnibus didn't pause the EU AI Act

The Digital Omnibus didn't pause the EU AI Act. It pushed the Annex III high-risk duties back to 2 December 2027, but other rules already apply and can be enforced now. Today's angle is for financial institutions.

Regulation (EU) 2026/1744, the Omnibus, amended Article 113 of the AI Act, the article that sets when each part applies. The core high-risk obligations in Chapter III now apply from 2 December 2027 to the high-risk uses listed in Annex III, which include creditworthiness and credit-scoring AI. High-risk AI built into products that fall under EU product-safety law (Annex I) follows later, from 2 August 2028.

That extra time is real, but it isn't a freeze. The bans on prohibited practices and the duty to make sure staff understand AI (AI literacy) have applied since February 2025. Since 2 August 2026, the Article 50(1) transparency rule has also applied. It requires AI that talks directly with people, such as a customer-service chatbot or robo-adviser, to tell them they're dealing with AI.

In practice, keep building your high-risk inventory and compliance plan against the December 2027 date. At the same time, check that your chatbots disclose they're AI and that nothing you run falls under a prohibited practice, since those rules are already live.