European companies hate the EU’s new AI rules — here’s why

Artificial intelligence sure is keeping the post office busy. After a recent flurry of open letters about runaway AI, unregulated AI, and apocalyptic AI, another missive arrived on the EU’s doorstep today.

In this case, however, the signatories have raised a contrary concern. Rather than call for more rules, they fear there will soon be too many.

Their target is the impending AI Act. Billed as the world’s first comprehensive legislation for the tech, the new rules are trying to walk the fine line between ensuring safety and supporting innovation. The new letter, signed by executives at some of Europe’s biggest companies, warns that they’re losing balance.

“The draft legislation would jeopardise Europe’s competitiveness and technological sovereignty without effectively tackling the challenges we are and will be facing,” said the letter, sent to the European Parliament, Commission, and member states.

“States with the most powerful models will have a decisive competitive advantage.

The signatories encompass bigwigs at corporate giants including Heineken, Carrefour, and Renault, as well as leaders at tech firms such as Ubisoft, TomTom, and Mistral AI. They say the AI Act will lead to companies leaving the bloc, investors redirecting cash, and stunted development in Europe.

One of their chief concerns stems from a recent change to the rules. On June 14, the European Parliament added new requirements on generative AI tools, such as ChatGPT, which would provoke “disproportionate” compliance costs and liability risks, according to the signatories.

They warn that this will push Europe further behind the US in AI development. This impact, they continue, will extend from the economy to culture, as large language models will be embedded into everything from search engines to digital assistants.

“States with the most powerful large language models will have a decisive competitive advantage… Europe cannot afford to stay on the sidelines,” said the letter.

Bending AI rules

In addition to raising the complaints, the business leaders proposed some solutions.

Their principal suggestion is restricting EU laws to broad principles in a risk-based approach, which would be implemented by a dedicated body and adaptable to new advances and risks. Such a process, they said, should be developed in dialogue with the economy.

The signatories also expressed support for some aspects of the AI Act. Specifically, they endorsed mandatory safety testing for new systems, standard labelling of AI-generated content, and a duty of care in model development.

Those overtures haven’t won over the lawmakers, however. Dragos Tudorache, who co-led the drafting of the AI Act, promptly rebuffed the letter.

“I am convinced they have not carefully read the text but have rather reacted on the stimulus of a few who have a vested interest in this topic,” he told Reuters.

On the plus side for the authors, there’s still time to write plenty more letters. The AI Act isn’t expected to come into force before 2026.