France and Germany may have found a sweet spot amid the European Union’s artificial intelligence rules. As expected, the bloc will require general-purpose AI players to report how they use data, train their models and respect copyright laws, but there’s a twist. Technology deemed to pose a “systemic risk” – to be determined by how much computing power it uses – will face additional requirements in a move that looks handy for homegrown players like France’s Mistral and Germany’s Aleph Alpha.
The computing heft of U.S. group OpenAI’s GPT-4 means it may be the only initial firm caught in the toughest enforcement net. At the other end, smaller European startups that will now need to do their own extensive due diligence in addition to developing their own new tech may lose out, potentially hurting competition. The relative winners are established but not yet massive homegrown players like Mistral, which said on Monday it raised 385 million euros in its second funding round.
As one of the first main jurisdictions to try to regulate a rapidly evolving and vital industry, Europe seemed to have already decided to err on the side of acting too soon rather than leaving room for innovation. Other regions may take a different view. For now, one upshot is that it creates some breathing space for European companies seen as most likely to succeed. (By Rebecca Christie)
Follow @Breakingviews on X
Capital Calls – More concise insights on global finance:
Anglo American’s reshuffle can go further read more
KKR’s smart meter swoop is well-timed read more
BAT’s smokeless future is clouded in smoke read more
Nio is on the long road to trimming losses read more
AI Davids ride coattails of industry Goliaths read more
