Big Tech may be able to breathe easier over European Union (EU) regulation.
The latest draft of provisions for the generally-strict EU AI Act published Tuesday contains softer language, indicating less-than-debilitating penalties for large U.S. tech companies. At the same time, Apple Inc. and Meta Platforms Inc. are likely to face modest fines, if any, for allegedly violating the Digital Markets Act (DMA) aimed at reeling in their power.
A third draft of the EU AI Act’s Code of Practice, presented as a “more streamlined structure with refined commitments and measures”, is packed with mediated language such as “reasonable measures” and “best efforts” when describing compliance with the law as it applies to data acquired for AI model training and models’ treatment of copyrighted content.
Gone from an earlier version of the code are strict requirements for providers of general purpose AI (GPAI) models to provide a single point of contact for handling rightsholders’ complaints “directly and rapidly.” The newest version also suggests GPAIs can decline to act on copyright complaints by rightsholders if they are “manifestly unfounded or excessive, in particular because of their repetitive character.”
Meanwhile, Apple and Meta appeared to have dodged a major regulatory bullet and resulting hefty fines for potentially breaching DMA after both companies had been in the European Commission’s crosshairs for months.
European antitrust authorities are now focused on making Apple and Meta comply with the law rather than face fines of up to 10% of their global annual sales, according to a report from Reuters. A final decision on the magnitude of the fines, if any, has not yet to be determined, the report said.
The watered-down provisions and lighter penalties against Apple and Meta come amid intensifying pressure from the U.S. government for aggressive European antitrust busters to back off. At an AI summit in Paris in February, Vice President JD Vance dismissed the need to regulate in Europe and instead urged government officials to embrace the “AI opportunity.”
Shortly after, the European bloc ditched the AI Liability Directive, a key safety initiative.
Last month, President Donald Trump intensified the rhetoric in a memorandum that threatened to levy tariffs against countries that fine U.S. companies.
Not all European regulators are backing down, however.
On Tuesday, Spain’s government approved a bill imposing fines of up to 35 million euros ($38.2 million) or 7% of global annual revenue on companies that use content-generated by AI without properly labelling it in a bid to curb the use of deepfakes.
The bill adopts guidelines from the EU AI Act.
“Regulation should be a guardrail, not a roadblock. Right now, European enterprises are caught in a tough spot. AI is evolving at lightning speed, and the rules keep shifting just as fast. Now we’re seeing policymakers soften their stance, which shows they recognize the risk of overregulation, but it also creates uncertainty. If the rules keep bending, companies won’t know what to prepare for,” Acrolinx CEO Matt Blumberg said in an email.
