"Artificial intelligence holds a huge, huge promise of incredible opportunities, but it also poses risks to our society, our economy, and our national security," Obama said.

Alongside executives from Amazon, Anthropic, Google, Inflection, Meta, Microsoft and OpenAI, Biden said that "the group here (present) will be critical to guiding this innovation responsibly and safely."

While these seven companies have committed to "three principles that must be fundamental to the development of artificial intelligence", namely safety, security and trust, Joe Biden praised their commitment to "push responsible innovation".

"We'll see more technological change in the next 10 years or even in the next few years than we've seen in the last 50 years. It was an amazing revelation for me," the Democratic president said.

2024 election

But fears about this powerful technology are intensifying, from dangers for consumers (fraud) and citizens (disinformation) to the risk of losing many jobs.

In concrete terms, these seven companies have promised to test their IT programs internally and externally before their launch. They also committed to investing in cybersecurity and sharing relevant information about their tools, including possible vulnerabilities, with authorities and researchers.

They must also "develop robust techniques to ensure that users know when content has been generated by AI, such as a watermark marking system," the White House added.

So far, the main companies involved have been rather reluctant to add such brands to the content created with their programs.

Fake photographs and deepfake montages have been around for years, but generative AI, which can produce text and images on simple queries in everyday language, raises fears of a flood of fake content online.

These can be used to manufacture ultra-credible scams or to manipulate public opinion. A particularly worrying prospect in the run-up to the 2024 US presidential election.

"Operate all levers"

The support committee of Ron DeSantis, Donald Trump's great rival in the race for the Republican Party's nomination, has already used in a commercial a fake voice generated by artificial intelligence and resembling that of the former president.

"We must use all the levers of the federal government to oversee" this sector and "work" with Congress to legislate, White House Chief of Staff Jeff Zients told the Axios news site.

He added that legislation is needed to have expertise within the federal government and to "have the regulatory authority to hold the private sector accountable for its actions."

Current political tensions in Congress -- where Republicans hold the House -- make new AI laws unlikely in the immediate future, but the government has said it is working on an executive order on AI security.

The industry commitments secured by the Biden administration are "an important first step," according to Paul Barrett, deputy director of New York University's Center for Business and Human Rights.

"But since they are not subject to punishment for non-compliance, it is vital that Congress votes on laws quickly," he added.

The White House also said it was working with allies abroad to seek "a strong international framework to govern the development and use of AI" around the world.

The topic was at the forefront of the G7 in Japan in May, and Britain is preparing to host an international summit on AI, probably next autumn.

© 2023 AFP