Artificial intelligence is advancing faster than anyone was prepared forand its starting to scare people. Now, the chiefs of two tech companies that are front-runners in A.I. are sharing the same messagethat governments should regulate A.I. so it doesnt get out of hand.
On Monday, top leaders of OpenAI, the maker of buzzy A.I. chatbot ChatGPT, said governments should work together to manage the risk of superintelligence, or advanced development of A.I. systems.
We can have a dramatically more prosperous future; but we have to manage risk to get there, OpenAI CEO Sam Altman wrote in a blogpost.
He wasnt the only one calling for more oversight of A.I. Alphabet and Google CEO Sundar Pichai also proposed that governments should be more involved in its regulation.
AI needs to be regulated in a way that balances innovation and potential harms, Pichai wrote in the Financial Times on Monday. I still believe AI is too important not to regulate, and too important not to regulate well.
Pichai suggested that governments, experts, academics, and the public all be part of the discussion when developing policies that ensure the safety of A.I. tools. The Google chief also said countries should work together to create robust rules.
Increased international co-operation will be key, Pichai wrote, adding that the U.S. and Europe must work together on future regulation in the A.I. space.
In his blogpost, Altman echoed the idea of better coordination on the safe development of A.I. rather than multiple groups and countries working separately. He said that one way to do this was by getting major governments around the world to set up a project that current A.I. efforts could become part of.
His other alternative was for a high-level governance body, akin to the United Nations International Atomic Energy Agency (IAEA) that oversees the use of nuclear power.
We are likely to eventually need something like an IAEA for superintelligence efforts, Altman wrote, adding that significant projects should be subject to an international authority that can inspect systems, require audits, test for compliance with safety standards, place restrictions on degrees of deployment and levels of security.
Another common thread from the two CEOs was the belief about the revolutionary impact of A.I. on human society. Altman said that superintelligence will be more powerful than other technologies humanity has had to contend with in the past, while Pichai repeated his famous proclamationthat A.I. is the most profound technology humanity is working on.
OpenAI and Google did not immediately return Fortunes request for comment.
As tech CEOs call for greater government involvement in regulations, some others argue that government regulation of A.I. will hamper innovation and that companies should regulate themselves.
My concern with any kind of premature regulation, especially from the government, is its always written in a restrictive way, former Google CEO Eric Schmidt told NBC News this month. What Id much rather do is have an agreement among the key players that we will not have a race to the bottom.
Growing calls for regulations
A.I.s potential risks are getting a lot of attention as the technology improves. In a Stanford University study published earlier this year, 36% of the experts surveyed acknowledged that A.I. would be ground-breakingbut said its decisions could lead to a nuclear-level catastrophe. Such tools could also be misused for nefarious aims and are often biased, directors at the universitys Institute for Human-Centered A.I. noted. The report also highlighted concerns that top companies could wind up with the most control over A.I.s future.
AI is increasingly defined by the actions of a small set of private sector actors, rather than a broader range of societal actors, the centers directors wrote.
Those fears have been voiced at the government level, too. Federal Trade Commission Chair Lina Khan, one of key voices in the discussion about anti-competitive practices, warned that A.I. could benefit only powerful actors if insufficiently regulated.
A handful of powerful businesses control the necessary raw materials that startups and other companies rely on to develop and deploy A.I. tools, Khan wrote in the New York Times this month. This means that control of A.I. could be concentrated with a few players, leading to the tech being trained on huge troves of data in ways that are largely unchecked.
Other A.I. experts, including Geoffrey Hinton, the so-called the Godfather of A.I. for his pioneering work in the field, have pointed out the technologys risks. Hinton, who quit his job at Google earlier this month, said he regretted his lifes work because of A.I.s potential dangers if put in the hands of bad actors. He also said that companies should only develop new A.I. if they are prepared for what it can do.
I dont think they should scale this up more until they have understood whether they can control it, Hinton said, referring to tech companies leading in the A.I. arms race.
Lawmakers have begun discussing regulating A.I. Last month, the Biden administration said it would seek public comments about possible rules. And Senate Majority leader Chuck Schumer, D-N.Y., is working with A.I. experts to create new rules and has already released a general framework, the Associated Press reported.
Responsible AI systems could bring enormous benefits, but only if we address their potential consequences and harms. For these systems to reach their full potential, companies and consumers need to be able to trust them, the National Telecommunications and Information Administrations Alan Davidson said in April, according to Reuters.
