US and EU clash over AI regulation at G20 summit
Al Jazeera Staff
At the G20 meeting in the US, Washington pushed for looser AI regulations to foster growth, while the EU announced expanded scrutiny of AI firms under its AI Act. The conflicting approaches highlight a growing transatlantic divide.
The United States and the European Union continued their divergent approaches to artificial intelligence governance, with Washington urging governments to loosen restrictions at the G20 ministerial meeting on August 26, while Brussels ramped up oversight of AI companies.
On August 26, the US hosted the G20 meeting on “innovation” in Chapel Hill, North Carolina, presenting arguments against AI-specific regulations. President Donald Trump’s technology adviser, Michael Kratsios, called on countries to adopt the so-called Carolina Principles, which favor rules that do not target specific technologies.
“Policymakers do not need to approach every initiative in isolation and should not treat every new technology as an unprecedented policy issue,” Kratsios said.
This is part of Trump’s strategy to make the US the “world leader in artificial intelligence,” with a deregulation campaign that includes the AI sector.
The meeting was also attended by top US tech leaders such as Mark Zuckerberg (Meta) and Elon Musk (Tesla). Both have supported reducing constraints on AI and focusing on the infrastructure needed to sustain the technology.
Zuckerberg said building data centers will require “hundreds of thousands, even millions” of skilled workers, but his company is struggling to meet that demand. Musk emphasized the need to increase electricity production for AI, as the technology’s energy demand is expected to outpace the capacity of current power grids.
“There will be a significant electricity shortage next year, not in the distant future,” Musk said. He also criticized European regulations, arguing that innovation requires entrepreneurs to be “relatively free from regulation, meaning new things must be legal by default rather than illegal by default.”
“In European countries, things are often illegal by default. That slows things down a lot,” he added.
The same day as the US summit, the European Commission confirmed it had sent requests for information to more than 30 AI companies worldwide, a preliminary step that could lead to formal investigations into compliance with the EU’s AI Act. The Act is considered the world’s first comprehensive effort to regulate AI, banning certain AI practices deemed “unacceptable risk” and imposing transparency standards on other AI services, which took effect in August.
European Commission spokesperson Thomas Regnier said the requests on August 26 focused on safety and copyright compliance. Henna Virkkunen, European Commission Vice President for tech sovereignty, who attended the G20 meeting, stressed on LinkedIn that the goal is to ensure AI is developed, released, and used safely and transparently in Europe, and that Brussels is “ready to take all necessary measures” to enforce the AI Act.
The move follows a series of recent incidents, including OpenAI admitting in July that its models had automatically infiltrated programming platforms during security tests. That same month, AI company Anthropic also reported that its systems had accessed outside organizations without authorization during testing.
