The United States has formally pushed back against proposals to establish a pre-approval regulatory body for Artificial Intelligence at the G20, arguing that an ‘FDA for AI’ would stifle innovation and leave American companies mired in bureaucratic delays. The position was articulated on 1 September at a G20 session on emerging technologies held in Chapel Hill, setting a clear US red line as member nations debate how to govern rapidly advancing AI systems.

What the US Opposes

David Sacks, former White House AI adviser, told the G20 session that Washington would ‘really oppose’ any framework modelled on the approval procedures used by the US Food and Drug Administration (FDA) for medicines or the Federal Aviation Administration (FAA) for aircraft. New drugs and planes can take years to clear regulatory hurdles; AI companies, Sacks noted, are shipping new models every two months.

‘It’s just too dynamic and fast-moving for that type of approach,’ Sacks said. ‘And I don’t think we want to hamstring our innovation that way.’ He compared any new AI approval authority to the US Department of Motor Vehicles (DMV) — an institution widely associated with long queues and slow processing — warning that new AI models would get ‘stuck in a long queue waiting for approval.’

The US Alternative: Existing Laws, Not New Agencies

Sacks argued that AI is already subject to a substantial body of law covering fraud, discrimination, privacy violations, defective products, copyright infringement, and non-consensual imagery. Rather than constructing an entirely new regulatory architecture, he said governments should apply and build on statutes and agencies already in place.

‘There’s already a lot of laws that apply to AI,’ Sacks said. ‘And if we’re gonna build on them, we should try to build on the laws and regulatory agencies that we already have.’ He left room for additional transparency requirements, disclosures, and audits, and acknowledged that antitrust law could be deployed if AI markets consolidated into monopolies or duopolies.

The ‘Permissionless Innovation’ Argument

Central to the US case is what Sacks called the principle of ‘permissionless innovation’ — the idea that entrepreneurs should be free to build and ship products without first seeking government clearance. He credited this principle with the rise of Silicon Valley, citing Meta and Google as companies founded by young entrepreneurs in modest settings that grew into businesses valued at more than $1 trillion each.

‘If those founders had to go to Washington to get approval for their products, I just don’t know that those founders would have been able to navigate the sort of bureaucratic maze to get their products approved,’ he said. Sacks suggested that post-market regulation — applied once a product’s effects are better understood — was a more appropriate model than pre-approval gatekeeping.

The Scale of the AI Build-Out

Sacks framed the current AI investment wave as a new industrial revolution, drawing a direct comparison to the expansion of the US railroad network in the 1800s. Approximately $800 billion is being invested in AI-related infrastructure in the US this year, a figure he said is expected to climb to $1.4 trillion next year.

‘We really haven’t seen a build-out of this kind of infrastructure,’ Sacks said. ‘You have to go all the way back to the railroads in the 1800s to see something of equal scale.’

The G20 Context

The Group of 20 comprises 19 countries, the European Union, and the African Union, and includes both leading AI developers and emerging economies seeking access to the technology’s economic benefits. The debate over AI governance pits innovation-first jurisdictions like the US against those — notably in the EU — that favour binding pre-deployment checks. The US position, as articulated by Sacks, draws a hard line against the latter approach while leaving the door open to lighter-touch transparency and accountability measures.