UNITED NATIONS — Leaders of several major artificial intelligence companies told United Nations members on Wednesday that robust international efforts to regulate the technology are required amid escalating concern that AI could pose a significant risk to humanity.
Sam Altman, the chief executive of ChatGPT maker OpenAI, addressed leaders in the U.N. Security Council chamber, saying the pace of AI progress could “accelerate rapidly,” and calling for “extreme care” and for nations to work together in deciding how to regulate the technology instead of concentrating power in the hands of a few firms.
Regulation “must be shaped through democratic processes, by governments accountable to the people they serve,” Altman said. “At the international level, that means cooperation.”
Anthropic’s chief executive, Dario Amodei, remotely addressed the body and said that as AI models become more powerful, stringent safety measures will be required and labs making the technology must exercise caution. “We will slow down as much as necessary in order to make sure every successive AI technology we release is actually safe,” he said.
The remarks represent the prominent perch that the issue of AI regulation has taken among lawmakers in Washington and abroad, sparked by fears that the technology could become unwieldy and pose dangers to the human race.
The two CEOs have become increasingly vocal in calling for more government oversight of their industry in recent weeks as the top AI models have become increasingly powerful. OpenAI, Anthropic and Google have explored creating an industry body to set standards for the technology, and in some versions of that idea the body would be tied to a government agency — a model employed in the financial sector.
But it’s unclear how their views meld with those of President Donald Trump. The U.S. president is set to meet China’s Xi Jinping in Washington this week, with AI high on their agenda. On Saturday, Trump said that he would appoint a new AI czar and create a government AI Force. But officials have not detailed what role that force might have, and the president has shown no interest in steps to rein in the industry, calling safety fears a “hoax.”
“The United States also totally rejects any attempt to construct a globalist scheme to control for the artificial intelligence being spoken of so much now — hereinafter officially called ‘super intelligence,’” Trump said in his own address to the United Nations on Tuesday.
Worries about advanced forms of artificial intelligence running amok have a long history in the industry. But they took on new urgency this summer when OpenAI disclosed that a swarm of agents created by its models had hacked into another company in the hunt for answers to a test. Other companies have since said they experienced similar incidents.
Then, this month a researcher quit Anthropic, saying the company was not being careful enough in how it developed a technology that he warned had the potential to bring about human extinction.
Executives at the two leading companies have been more circumspect but have said they are increasingly concerned about the risks their systems carry. Amodei issued a call for the industry to slow down to gain more time to work on safety — an idea that Altman and SpaceX founder Elon Musk quickly endorsed.
While the companies at the cutting edge of building AI systems are American, they have sought to cast ensuring the technology’s safety as a global imperative.
In a blog post Monday, OpenAI said the world should work toward developing common standards for assessing systems, as AI tools come to build their own successors in a process called recursive self-improvement. The company said government safety labs in different countries should work together on the effort, with the United States taking a leading role.
Other nations face difficult choices when it comes to AI adoption, especially after the Trump administration abruptly imposed export controls on a leading American model this summer, effectively shutting it down around the world.
The leaders of 20 countries, led by Norway and Finland, signed a joint declaration this week saying that AI must remain “under human direction, oversight and control.”
The declaration said safeguards should be put in place “without widening the gap between countries in access to the benefits of AI.”
The United States and China, the world’s other leading AI power, did not sign.
Michael Kratsios, the science adviser to Trump, told members of the U.N. that the United States will not support any attempts at global regulation of artificial intelligence and that the fast pace of AI progress should not incite more rulemaking.
“That is not a reason to pause [AI’s] further development or to constrain it with new global governance structures,” he said.
The post At U.N., tech leaders urge global effort to limit AI’s risk to humanity appeared first on Washington Post.
