Trending Now
We have updated our Privacy Policy and Terms of Use for Eurasia Group and its affiliates, including GZERO Media, to clarify the types of data we collect, how we collect it, how we use data and with whom we share data. By using our website you consent to our Terms and Conditions and Privacy Policy, including the transfer of your personal data to the United States from your country of residence, and our use of cookies described in our Cookie Policy.
{{ subpage.title }}
FILE PHOTO: California Governor Gavin Newsom (D) reacts as he speaks to the members of the press on the day of the first presidential debate hosted by CNN in Atlanta, Georgia, U.S., June 27, 2024.
Gov. Gavin Newsom vetoes California’s AI safety bill
California Gov. Gavin Newsom on Sunday vetoed the Safe and Secure Innovation for Frontier Artificial Intelligence Models Act, or SB 1047, the AI safety bill passed by the state’s legislature in August.
Newsom has signed other AI-related bills into law, such as two recent measures protecting performers from AI deepfakes of their likenesses, but vetoed this one over concerns about the focus of the would-be law.
“By focusing only on the most expensive and large-scale models, SB 1047 establishes a regulatory framework that could give the public a false sense of security about controlling this fast-moving technology,” Newsom wrote in a letter on Sept. 29. “Smaller, specialized models may emerge as equally or even more dangerous than the models targeted by SB 1047 — at the potential expense of curtailing the very innovation that fuels advancement in favor of the public good.”
Democratic state Sen. Scott Wiener, who sponsored the bill, called the veto a “setback for everyone who believes in oversight of massive corporations that are making critical decisions that affect the safety and welfare of the public and the future of the planet.” Wiener hasn’t disclosed the next steps but vowed to continue pushing the envelope on AI regulation in the state. “California will continue to lead in that conversation — we are not going anywhere.”Commerce Secretary Gina Raimondo arrives to a Senate Appropriations Subcommittee on Commerce, Justice, Science, and Related Agencies hearing on expanding broadband access on Capitol Hill in Washington, D.C., U.S. February 1, 2022.
National safety institutes — assemble!
The Biden administration announced that it will host a global safety summit on artificial intelligence on Nov. 20-21 in San Francisco. The International Network of AI Safety Institutes, which was formed at the AI Safety Summit in Seoul in May, will bring together safety experts from each member country’s AI safety institute. The current member countries are Australia, Canada, the European Union, France, Japan, Kenya, Singapore, South Korea, the United Kingdom, and the United States.
The aim? “Strengthening international collaboration on AI safety is critical to harnessing AI technology to solve the world’s greatest challenges,” Secretary of State Antony Blinken said in a statement.
Commerce Secretary Gina Raimondo, co-hosting the event with Blinken, said that the US is committed to “pulling every lever” on AI regulation. “That includes close, thoughtful coordination with our allies and like-minded partners.”What do Democrats want for AI?
At last week’s Democratic National Convention, the Democratic Party and its newly minted presidential candidate, Vice President Kamala Harris, made little reference to technology policy or artificial intelligence. But the party’s platform and a few key mentions at the DNC show how a Harris administration would handle AI.
In the official party platform, there are three mentions of AI: First, it says Democrats will support historic federal investments in research and development, break “new frontiers of science,” and create jobs in artificial intelligence among other sectors. It also says it will invest in “technology and forces that meet the threats of the future,” including artificial intelligence and unmanned systems.
Lastly, the Dems’ platform calls for regulation to bridge “the gap between the pace of innovation and the development of rules of the road governing the most consequential domains of technology.”
“Democrats will avoid a race to the bottom, where countries hostile to democratic values shape our future,” it notes.
Harris echoed that final point in her DNC keynote address. “I will make sure that we lead the world into the future on space and artificial intelligence,” she said. “That America, not China, wins the competition for the 21st century, and that we strengthen, not abdicate our global leadership.”
The Republican Party platform, by contrast, promises to repeal Biden’s 2023 executive order on AI, calling it “dangerous,” hindering innovation, and imposing “radical left-wing ideas” on the technology. “In its place, Republicans support AI development rooted in free speech and human flourishing,” it says. (The platform doesn’t go into specifics about how the executive order is harmful or what a free speech-oriented AI policy would entail.) In his RNC address, Donald Trump didn’t mention artificial intelligence or tech policy but talked at length about beating back China economically.
GZERO asked Don Beyer, the Virginia Democratic congressman going back to school to study artificial intelligence, what he thought of his party’s platform and Harris’ remarks on AI. Beyer said that Harris has struck the right balance between promoting American competitiveness and outlining guardrails to minimize the technology’s risks. “The vice president has been personally involved in many of the administration’s efforts to ensure American leadership in AI, from establishing the US AI Safety Institute to launching new philanthropic initiatives for public interest AI, and I expect her future administration to continue that leadership,” he said.California wants to prevent an AI “catastrophe”
The Golden State may be close to passing AI safety regulation — and Silicon Valley isn’t pleased.
The proposed AI safety bill, SB 1047, also known as the Safe and Secure Innovation for Frontier Artificial Intelligence Models Act, aims to establish “common sense safety standards” for powerful AI models.
The bill would require companies developing high-powered AI models to implement safety measures, conduct rigorous testing, and provide assurances against "critical harms," such as the use of models to execute mass-casualty events and cyberattacks that lead to $500 million in damages. It warns that the California attorney general can take civil action against violators, though rules would only apply to models that cost $100 million to train and pass a certain computing threshold.
A group of prominent academics, including AI pioneers Geoffrey Hinton and Yoshua Bengio,published a letter last week to California’s political leaders supporting the bill. “There are fewer regulations on AI systems that could pose catastrophic risks than on sandwich shops or hairdressers,“ they wrote, saying that regulations are necessary not only to rein in the potential harms of AI but also to restore public confidence in the emerging technology.
Critics, including many in Silicon Valley, argue the bill is overly vague and could stifle innovation. In June, the influential startup incubator Y Combinator, wrote a public letter outlining its concerns. It said that liability should lie with those who abuse AI tools, not developers, that the threshold for inclusion under the law is arbitrary, and that a requirement that developers include a “kill switch” allowing them to turn off the model would be a “de facto ban on open-source AI development.”
Steven Tiell, a nonresident senior fellow with the Atlantic Council's GeoTech Center, thinks the bill is “a good start” but points to “some pitfalls.” He appreciates that it only applies to the largest models but has concerns about the bill’s approach to “full shutdown” capabilities – aka the kill switch.
“The way SB 1047 talks about the ability for a ‘full shutdown’ of a model – and derivative models – seems to assume foundation models would have some ability to control derivative models,” Tiell says. He warned this could “materially impact the commercial viability of foundation models across wide swaths of the industry.”
Hayley Tsukayama, associate director of legislative activism at the Electronic Frontier Foundation, acknowledges the tech industry’s concerns. “AI is changing rapidly, so it’s hard to know whether — even with the flexibility in the bill — the regulation it’s proposing will age well with the industry,” she says.
“The whole idea of open-source is that you’re making a tool for people to use as they see fit,” she says, emphasizing the burden on open-source developers. “And it’s both harder to make that assurance and also less likely that you’ll be able to deal with penalties in the bill because open-source projects are often less funded and less able to spend money on compliance.”
State Sen. Scott Wiener, the bill’s sponsor, told Bloomberg he’s heard industry criticisms and made adjustments to its language to clarify that open-source developers aren’t entirely liable for all the ways their models are adapted, but he stood by the bill’s intentions. “I’m a strong supporter of AI. I’m a strong supporter of open source. I’m not looking in any way to impede that innovation,” Wiener said. “But I think it’s important, as these developments happen, for people to be mindful of safety.” Spokespeople for Wiener did not respond to GZERO’s request for comment.
In the past few months, Utah and Colorado have passed their own AI laws, but they’ve both focused on consumer protection rather than liability for catastrophic results of the technology. California, which houses many of the biggest companies in AI, has broader ambitions. But while California has been able to lead the nation — and the federal government on data privacy — it might need industry support to get its AI bill fully approved in the legislature and signed into law. California’s Senate passed the bill last month, and the Assembly is set to vote on it before the end of August.
California Gov. Gavin Newsom hasn’t signaled whether or not he’ll sign the bill should it pass both houses of the legislature, but in May, he publicly warned against over-regulating AI and ceding America’s advantage to rival nations: “If we over-regulate, if we overindulge, if we chase the shiny object, we could put ourselves in a perilous position.”The FEC kicks AI down the road
That means that the job of keeping deepfakes out of political ads will largely fall to tech platforms and AI developers. Tech companies signed an agreement at the Munich Security Conference in February, vowing to take “reasonable precautions” to prevent their AI tools from being used to disrupt elections. It’s also a task that could potentially fall to broadcasters. The Federal Communications Commission is still considering new rules for AI-generated content in political ads on broadcast television and radio stations. That’s caused tension between the two agencies too: The FEC doesn’t believe the FCC has the statutory authority to act, but the FCC maintains that it does.
After a deepfake version of Joe Biden’s voice was used in a robocall in the run-up to the New Hampshire Democratic primary, intended to trick voters into staying home, the FCC asserted that AI-generated robocalls were illegal under existing law. But time is ticking for further action since other AI-manipulated media may not be covered currently under the law. At this point, it seems likely that serious regulation from either agency might only come after Donald Trump and Kamala Harris square off in November — and perhaps only if Harris wins, as another Trump presidency might mean a further rollback of election rules.An Apple logo is pictured in an Apple store in Paris, France.
Apple signs Joe Biden’s pledge
Apple signed on to the Biden administration’s voluntary pledge for artificial intelligence companies on July 26.
President Joe Biden and Vice President Kamala Harrisfirst announced that they secured commitments from seven major AI developers — Amazon, Anthropic, Google, Inflection, Meta, Microsoft, and OpenAI — a year ago in what the administration says laid the groundwork for its executive order on AI adopted in October. The voluntary commitments included safety testing, information sharing on safety risks (with government, academia, and civil society groups), cybersecurity investments, watermarking systems AI-generated content, and a general agreement to “develop and deploy advanced AI systems to help address society’s greatest challenges.”
Until now, Apple wasn’t on the list. Now, as Apple prepares to release new AI-enabled iPhones (powered by OpenAI’s systems as well as its own), the Cupertino-based tech giant is playing nice with the Biden administration, signaling that they’ll be a responsible actor, even without formal legislation on the books.
What Sam Altman wants from Washington
Altman’s argument is not new, but his policy prescriptions are more detailed than before. In addition to the general undertone that Washington should trust the AI industry to regulate itself,the OpenAI chief calls for improved cybersecurity measures, investment in infrastructure, and new models for global AI governance. He wants additional security and funding for data centers, for instance, and says doing this will create jobs around the country. He also urges the use of additional export controls and foreign investment rules to keep the AI industry in US control, and outlines potentially global governance structures to oversee the development of AI.
We’ve heard Altman’s call for self-regulation and industry-friendly policies before — he has become something of a chief lobbyist for the AI industry over the past two years. His framing of AI development as a national security imperative echoes a familiar strategy used by emerging tech sectors to garner government support and funding.
Scott Bade, a senior geotechnology analyst at Eurasia Group, says Altman wants to “position the AI sector as a national champion. Every emerging tech sector is doing this:
‘We’re essential to the future of US power [and] competitiveness [and] innovation so therefore [the US government] should subsidize us.’”
Moreover, Altman’s op-ed has notable omissions. AI researcher Melanie Mitchell, a professor at the Santa Fe Institute,points out on X that there’s no mention of the negative effects on the climate, seeing that AI requires immense amounts of electricity. She also highlights a crucial irony in Altman’s insistence to safeguard intellectual property: “He’s worrying about hackers stealing AI training data from AI companies like OpenAI, not about AI companies like OpenAI stealing training data from the people who created it!”
The timing of Altman’s op-ed is also intriguing. It comes as the US political landscape is shifting, with the upcoming presidential election no longer seen as a sure win for Republicans. The race between Kamala Harris and Donald Trump is now considered a toss-up, according to the latest polling since Harris entered the race a week and a half ago. This changing dynamic may explain why Altman is putting forward more concrete policy proposals now rather than counting on a more laissez-faire approach to come into power in January.
Harris is both comfortable with taking on Silicon Valley and advocating for US AI policy on a global stage, as we wrote in last week’s edition. Altman will want to make sure his voice — perhaps the loudest industry voice — gets heard no matter who is elected in November.A photo illustration of a smartphone displaying the NVIDIA Corporation stock price on the NASDAQ market, with an NVIDIA chip visible in the background.
A chip bottleneck
Margrethe Vestager, the European Union’s competition chief, has warned of a “huge bottleneck” involving Nvidia. The US semiconductor company plays a pivotal role in designing chips necessary for training and running artificial intelligence models and applications — good for 80% of the market. In recent months, Nvidia has become a $3.1 trillion company — now the third-most-valuable firm in the world behind only Microsoft and Apple.
It’s too much of a good thing: NVIDIA’s chips are so in demand that it can’t make enough for AI firms looking to train bigger and better models. EU regulators are starting to wonder whether that bottleneck raises concerns over fair competitive markets.
In Singapore, Vestager told Bloomberg that the EU’s watchdogs are asking preliminary questions of Nvidia but haven’t made up their mind about any further regulatory steps. Vestager said a robust secondary market could relieve competitive concerns, implying that as long as Nvidia respects smaller firms it should stay in antitrust regulators’ good graces. Meanwhile, AMD and Intel are looking to close the gap with Nvidia, something that intense regulatory scrutiny on the market leader might aid.