OpenAI's chief: regulate me, please

Hi, it’s Dina and Rachel on the West Coast. OpenAI’s Sam Altman told Congress he’s “eager” for artificial intelligence to be regulated. But [View in browser]( [Bloomberg]( Hi, it’s Dina and Rachel on the West Coast. OpenAI’s Sam Altman told Congress he’s “eager” for artificial intelligence to be regulated. But first... Today’s must-reads: • WeWork’s CEO [stepped down]( • Meta’s refunds are slow to [get to small advertisers]( • Alleged Russian hacker charged in [ransomware spree]( Altman lays out his case for regulation On Tuesday, [Sam Altman](bbg://people/profile/16408119), the 38-year old chief executive officer of OpenAI, [told US senators]( about a variety of areas where he’d support regulation, including preventing election misinformation, making AI-generated content clear to users — oh, and stopping people from using AI to kill us all. He also spoke about the promise of artificial intelligence and detailed its potential harms. “If this technology goes wrong, it can go quite wrong,” he said. One near-term “uh-oh” Altman sees on the horizon: AI systems that seek to influence voters ahead of the 2024 US elections. “It’s one of my areas of greatest concern — the general ability of these models to provide one-on-one interactive disinformation,” said Altman, speaking at a three-hour hearing before a Senate subcommittee. Who will impose and enforce these and other rules? Well, Altman expressed “enthusiastic” support for a new federal agency to examine the work of AI companies (another panelist, Christina Montgomery, the chief privacy and trust officer at IBM, disagreed, saying rules should come from agencies already in place). Altman also backed the idea of that agency issuing licenses for complying AI products and being empowered to yank a license if violations turn up later. Altman said he wants the US to lead the way, but make sure the impacts are global. One way the US could impose effective worldwide standards, the tech executive suggested, was via its control over the microprocessors that train and run AI systems. Those chips, called GPUs, are largely sold by one US company: Nvidia Corp. The supply is limited, and much of it is in the hands of other large US tech companies. Depending on how any legislation is written, following Altman’s lead runs the risk of cementing the power currently held by leaders in the AI industry — companies like OpenAI, Alphabet Inc.’s Google, Microsoft Corp. and Amazon.com Inc. — and [stifling innovation](from startups, particularly those connected to the bustling open-source software community. This might not be the aim, but it would come at an auspicious time for the current crop of major players; they are keenly aware of recent rapid advances in open source AI projects, including the development of increasingly capable large language models. That work has so much potential a senior software engineer at Google recently said in a leaked critique that the open source community is “eating our lunch.” Senator [Cory Booker](bbg://people/profile/4632738), a New Jersey Democrat, had some of the same concerns, asking Altman about “massive corporate concentration” in AI that is “really terrifying.” Altman agreed with lawmakers that any new legislation should not end up kneecapping startups and smaller efforts. “It’s important that any new approach, any new law does not stop the innovation from happening with smaller companies, open source models, researchers that are doing work at a smaller scale,” he said. “That’s a wonderful part of this ecosystem and of America and we don’t want to slow it down.” Instead he suggested a two-tier system — the easiest way is to impose categories based on how much computing power an AI system uses, with larger models getting the extra scrutiny. But Altman prefers a more complex — and in his view more accurate — way of deciding who gets tighter regulation: The idea would be to focus regulation on AI programs that boast specific and potentially harmful capabilities. What capabilities, you may ask? Well, Senator [Jon Ossoff](bbg://people/profile/20124689) of Georgia did just that, and after initially demurring, Altman listed models that persuade or manipulate a person’s beliefs or create novel biological agents. Some of the senators seemed to think highly of Altman’s point of view. Senator [John Kennedy](bbg://people/profile/3806782) of Louisiana worried whether “a berserk wing of the artificial intelligence community could intentionally or unintentionally use artificial intelligence to kill all of us and hurt us the entire time we are dying.” When Altman reiterated that we need an agency to regulate and license AI, Kennedy offered him the job of running it, but Altman turned it down. Putting aside the very significant question of whether it’s a good idea to ask leaders of an industry how to regulate that same industry (cough cough), at least there appeared to be bipartisan support for AI regulations. If you watched the hearing, you’d be forgiven for thinking something might happen soon. Color us unconvinced. —[Dina Bass](mailto:dbass2@bloomberg.net) and [Rachel Metz](mailto:rmetz17@bloomberg.net) The big story CNET journalists are seeking to unionize. They say AI [“threatens our jobs and reputations.”]( Get fully charged The US is pursuing criminal charges in a campaign to [stop Russia and China from acquiring sensitive technologies](. Twitter was sued for allegedly helping [silence critics of Saudi Arabia](. ServiceNow joined the AI bandwagon in a [partnership with Microsoft and OpenAI](. Israel is illegally storing millions of people’s photos, [an audit found](. More from Bloomberg Get Bloomberg Tech newsletters in your inbox: - [Cyber Bulletin]( for coverage of the shadow world of hackers and cyber-espionage - [Game On]( for reporting on the video game business - [Power On]( for Apple scoops, consumer tech news and more - [Screentime]( for a front-row seat to the collision of Hollywood and Silicon Valley - [Soundbite]( for reporting on podcasting, the music industry and audio trends - [Hyperdrive]( for expert insight into the future of cars Follow Us Like getting this newsletter? [Subscribe to Bloomberg.com]( for unlimited access to trusted, data-driven journalism and subscriber-only insights. Want to sponsor this newsletter? [Get in touch here](. You received this message because you are subscribed to Bloomberg's Bloomberg Tech Daily newsletter. If a friend forwarded you this message, [sign up here]( to get it in your inbox. [Unsubscribe]( [Bloomberg.com]( [Contact Us]( Bloomberg L.P. 731 Lexington Avenue, New York, NY 10022 [Ads Powered By Liveintent]( [Ad Choices](