The White House has stolen a march on Rishi Sunak’s AI safety summit by unveiling sweeping regulations aimed at protecting workers and citizens from threats posed by the technology.
US President Joe Biden will sign an executive order aiming to guide the development of the tech, requiring firms working on potentially dangerous models to share safety data with the government before their release.
It also aims to protect workers’ rights to address concerns that artificial intelligence will lead to redundancies, and ensure models are not trained in a way that exacerbates existing societal biases.
And in a bid to combat misinformation, such as from deepfake videos, the White House is working on “content authentication and watermarking” guidelines for content generated by AI.
White House deputy chief of staff Bruce Reed described it as the “strongest set of actions” any government has announced on AI.
Mr Biden is due to speak about the safeguards later on Monday, days before the UK’s safety summit.
Mr Sunak has invited world leaders and tech bosses to Britain for a two-day event this week, but Mr Biden isn’t attending and will send vice president Kamala Harris.
UK has ‘not at all’ been snubbed
Starting on Wednesday at Bletchley Park, home of Britain’s Second World War codebreakers, the summit will feature discussions about how the potential of AI can be harnessed while managing the risks.
Mr Sunak has been accused of focusing too much on the existential threats, like humanity losing control of the tech, rather than near-term challenges like its impact on jobs.
But he still hopes the summit will cement the UK’s status as a leading AI power, despite expected absentees like Canada’s Justin Trudeau, France’s Emmanuel Macron, and Germany’s Olaf Scholz.
Asked whether Mr Sunak feels he’s being snubbed, his spokesperson said: “No, not at all.”
They added: “We remain confident that we have brought together the right group of world experts in the AI space, leading businesses, and indeed world leaders and representatives who will be able to take on this vital issue.”
Read more:
Sunak walking treacherous tightrope over AI
Sunak vows to tackle fears around AI
Biden determined to ‘move fast’
The White House follows the EU and China in unveiling its own proposed AI regulation, while the G7 is expected to announce joint principles shortly.
Mr Biden is said to have told his staff of the need for urgency, having reportedly expressed regrets about how slow governments were to regulate social media.
Many of the world’s leading AI companies are headquartered in the US, including ChatGPT maker OpenAI.
Deputy chief of staff Mr Reed said the president is “impressed and alarmed” by the capability of AI.
“He saw fake AI images of himself, of his dog. He saw how it can make bad poetry,” he said.
“And he’s seen and heard the incredible and terrifying technology of voice cloning.”
Is AI an existential threat?
The role of Mission: Impossible
Mr Reed revealed that in addition to meetings with leading figures in the field, Mr Biden had even been faced with threats posed by AI in his downtime.
At his Camp David retreat, he sat down to watch the latest Mission: Impossible, which sees Tom Cruise face his greatest villain yet: a sentient rogue AI.
“If he hadn’t already been concerned about what could go wrong with AI before that movie,” said Mr Reed.
“He saw plenty more to worry about.”
