EU tech chief eyes AI Act amendments to create legal certainty

The European Commission is planning "targeted amendments” to the bloc’s artificial intelligence rulebook next week, Henna Virkkunen, the European Commissioner for Tech Sovereignty, Security and Democracy, told Euronews at the Web Summit tech conference in Lisbon.

The AI Act – rules that regulate artificial intelligence tools based on the risks they pose to society – began applying gradually last year.

However, the laws faced ongoing criticism by Big Tech companies, as well as the US administration led by Donald Trump, claiming that they stifle innovation.

“The next important part [of the AI Act entering into force] will be next August. And there we are really facing challenges because we don't have the [technical] standards yet - and they need to be ready one year before the next phase,” Virkkunen said on Tuesday.

“Now, we have to look at how we can create legal certainty for our industries, and that's something that we are now considering: how we can support our industries when we don't have the standards in place.”

Virkkunen added that the amendments to the AI Act - to be presented on 19 November - still need formal approval by the College of Commissioners as a whole.

She stopped short of saying how far-reaching those changes will be and whether they will include a formal pause of some of the law’s provisions. Virkkunen said that the Commission remains "very committed to the main principles [of the law]".

The so-called digital omnibus package, which is an effort by the Commission to cut red tape and make the lives of companies easier by reducing their administrative burden, will also include changes to the EU’s data policy and cybersecurity rules.

Pressure on changes to the AI Act

According to drafts of the plans that have been circulating, the simplification package could introduce a one-year grace period, meaning that national authorities can fine misuse only as of August 2027.

Earlier this year, CEOs of more than 40 European companies, including ASML, Philips, Siemens and Mistral AI, asked for a “two-year clock-stop” on the AI Act before key obligations enter into force.

The Commission has repeatedly said that it is not giving in to any external pressure concerning the possible delay of certain provisions.

Michael O'Flaherty, Human Rights Commissioner at the Council of Europe - Europe's leading human rights organisation - warned about the consequences of the simplification plans while speaking to Euronews at Web Summit.

"Let's be very careful not to discard the [laws'] core protective elements," O'Flaherty said.

"If there's a way to join up multiple regulations in a more efficient manner, fine, but let's not throw out the baby with the bathwater. Let's not give in to the very heavy tech lobby to make life less onerous for tech and, as a result, more risky for us," he said.