AI Law Draws Pushback From Big Brands in Europe

BRUSSELS—Regulators in Europe are racing to create the West’s first comprehensive set of rules for artificial intelligence. Some businesses say their plans go too far.

Executives from dozens of large European companies including engineering giant Siemens, Dutch brewer Heineken and French carmaker Renault issued an open letter Friday warning that the rules as drafted risk driving companies and capital away from the continent. Other signatories included leaders from a handful of European startups and a top AI researcher at Meta Platforms.

“The result would be a critical productivity gap” between the U.S. and Europe, the executives wrote.

The letter adds to a lobbying effort from business leaders, academics and AI companies over whether and how to set out rules for the powerful technology behind tools such as OpenAI’s ChatGPT and Google’s Bard.

European lawmakers are hammering out draft legislation that could require companies to design their AI models in a way that prevents them from creating illegal content. Companies might also be required to publish summaries of the copyrighted data used to train their models and to disclose whenever content is generated using AI.

Other provisions could include restrictions on some types of biometric surveillance and on the way images are gathered for facial-recognition databases.

Officials from the European Commission, the European Parliament and member states say they want to come up with a compromise agreement on the new rules before the end of this year.

Politicians, academics, activists and many tech executives have pushed for some form of AI regulation. But the breadth of the industries represented by the letter’s signatories—some far from Silicon Valley—demonstrates an unease among many companies about taking a prescriptive approach in reining in the technology. 

After ChatGPT and Bard were launched, companies have scrambled to learn how they can put such tools to use, including how they might lower costs.   

The business executives behind this week’s letter said they believe AI regulation is needed. They said their main concern is with the addition of new provisions focused on generative AI to the draft legal text, which they said were too rigid. They include broad rules for how AI systems are trained on very large data sets, and would apply regardless of how those models are used and the level of risk involved.

The executives said the law should use a risk-based approach and focus on broad principles that could be implemented by a regulatory body made up of experts.

Under the proposed rules, “companies developing and implementing such systems would face disproportionate compliance costs and disproportionate liability risks,” the executives said. 

“Such regulation could lead to highly innovative companies moving their activities abroad” and investors withdrawing capital from European AI development, they said.

EU lawmaker Dragos Tudorache, who co-led the institution’s work on the AI legislation, said he believes the legislation as drafted would solicit industry input in defining standards and in setting up governance, and provides a light regulatory regime. 

—Sam Schechner contributed to this article.

Write to Kim Mackrael at [email protected]