As Vice President Kamala Harris continues to coalesce support for her presidential run, The Tech Oversight Project is spotlighting her leadership on tech and AI safety. Late last year, our organization led a coalition of tech safety, labor, and competition advocates in thanking Vice President Harris for highlighting the risks and harms of AI affecting people’s lives right now. View the letter here.
“Vice President Kamala Harris has been a champion for AI safety and protecting Americans from the real-world harms that are happening right now. With the AI industry continuing to grow and shift, we need to strike the right balance between common sense guardrails and innovation. We applaud Vice President Harris for her leadership and look forward to working with her to protect consumers and help grow our economy,” said Sacha Haworth, Executive Director of the Tech Oversight Project.
⚡ FLASHBACK: Kamala Harris seizes agenda as Rishi Sunak’s AI summit kicks off
US vice president urged the international community to focus on the ‘full spectrum’ of artificial intelligence risks, including bias and discrimination.
By Eugene Daniels and Vincent Manancourt on 11/1/23
LONDON — Kamala just showed Rishi who’s boss.
As British Prime Minister Rishi Sunak’s showpiece artificial intelligence event kicked off in Bletchley Park on Wednesday, 50 miles south in the futuristic environs of the American Embassy in London, U.S. Vice President Kamala Harris laid out her vision for how the world should govern artificial intelligence.
It was a raw show of U.S. power on the emerging technology.
Armed with a freshly released White House Executive Order on AI and walking out to the piping trumpets of a brass band, Harris could point to comprehensive domestic action the Biden administration is taking, including new binding guardrails on the most advanced forms of the emerging tech which American companies dominate.
By contrast, for all the brouhaha around Sunak’s summit — and notable diplomatic breakthroughs with the Bletchley Declaration signed by Beijing, Washington and others — Britain is falling behind the likes of the EU, Canada and now the U.S. in bringing through concrete rules around AI.
Harris also urged the international community to focus on the “full spectrum” of artificial intelligence risks, including existing threats like bias and discrimination. It was a gentle rebuke to Sunak’s summit, which has courted controversy due to its laser focus on the unrealized existential risks of the tech.
“Let us be clear, there are additional threats that also demand our action. Threats that are currently causing harm, and which to many people also feel existential,” Harris said.
In her speech, the small differences between the countries’ approaches did not go unnoticed “She brought it down to a level for people to really think about, ‘OK, here’s how it could affect your day to day, or here’s some of the other implications that we’re not really concerned about,’” Toccara Baker, an American marketing executive living in the U.K., said after the speech.
Harris aides told POLITICO one of veep’s intentional goals was to tell the global community that America’s way of approaching AI should be the blueprint for every country, including the U.K. “You’re picking up what we’re putting down,” one White House official said.
Harris aides often remark the vice president “doesn’t like to come empty handed” on trips. Multiple officials said she pushed the administration to make sure she’d have announcements to make and a finished executive order to tout.
“She talked to people within the White House, to make sure that, ‘Hey, it’s June. We know we’re going to do an EO. We’ve now announced that we’re doing an EO. I want to see the timeline of how and how you are going to actually get that to occur as fast as possible,’” as one White House official put it.
Matt Clifford, Sunak’s lead official on the AI summit, pushed back on the idea that this week’s international meet-up wasn’t focused on near-term risks.
“I think the short-term risks, long-term risks distinction is overblown,” he told a pack of reporters at Bletchley Park. “We know right now that companies are training models, that in some cases are 50 times more powerful from the current models. These are near term risks.”
White House officials make clear, while trying to remain diplomatic, that Harris was intentional about adding that aspect of the speech.
“The U.S. and the U.K. are very close allies and partners on the full range of global issues, including AI. And there is a lot that we agree on. But I think we’ve been very clear and the vice president was very clear in her speech that there are certain issues that we believe the entire global community should work on together,” another U.S. official said.
Sunak also pushed back on claims Washington stole his thunder by issuing the executive order on AI.
“I think actually the opposite. I’m really pleased that the U.S. have taken this moment when we’re convening this summit for the vice president to give a really important speech on AI and for the president to release the executive order,” he said on the POLITICO Power Play podcast.
The U.S. also announced the creation of an American AI Safety Institute, first reported by POLITICO, which will work with Sunak’s own institute in preventing catastrophes stemming from AI.
But there are question marks over how the two institutes will collaborate once the U.S. throws its weight behind its own initiative, given Washington’s nervousness about national security, and the fact the vast majority of leading AI companies are in the United States.
For the moment, at least, Britain is relying on its head start in the area. Its Frontier AI Taskforce is already in talks with companies like Google DeepMind, Anthropic and OpenAI to get access to their models.
In private, however, AI companies fret about a proliferation of similar institutes around the world that want to get under the bonnet of their tech. Leading AI labs are currently working with Britain’s initiative on a voluntary basis, while the U.S. order places binding requirements on them to hand over information.
Ahead of a second day of the summit which will involve a smaller group of like-minded countries to discuss the national security aspects of AI, Sunak can reflect on a quietly successful first day at Bletchley — despite the side-eye from Harris in London.
In nothing short of a diplomatic coup for the Brits, Sunak got the U.S., EU and China to put pen to paper on a shared communiqué outlining the risks from the most advanced forms of the tech.
“I’m pleased that [China] has engaged,” Sunak told POLITICO. “It hasn’t happened for a long time that, you know, we’ve all signed the same bit of paper, that’s us, the Americans, Europeans in China, which I think is a good sign of progress.”
News that other capitals are to take the baton from Sunak on AI safety summits will be vindication for his decision to hold one. There are to be two subsequent such meet-ups: the first hosted by Korea in six months, the second one in a year’s time in France.