While President Biden was in San Francisco finalizing an agreement between the US and China to initiate a conversation on artificial intelligence and other matters, a prominent group of AI and military leaders convened in the Utah mountains for a covert event. In the span of three days last week, over 100 AI executives, insiders, venture capitalists, and government officials attended presentations on the AI arms race between the US and China, and deliberated on the implications of Biden’s comprehensive executive order on the industry. They explored the potential misuse of AI by malicious individuals and contemplated the future transformative developments in the field.
The AI Security Summit is hosted by Scale AI Inc., a startup that provides artificial intelligence training and data annotation services to clients ranging from OpenAI and General Motors Co. to the U.S. military. Only those invited had reason to know about it. The all-black, password-protected website of the summit gave little meaning. “Join the world’s brightest AI leaders and visionaries,” read a post on the site. “By invitation only.”
For the AI community, the summit provided an opportunity to discuss the uncertainties ahead — AI regulation, political conflicts and the technology itself — in a luxurious retreat far from the hearing rooms of Washington and the boardrooms of Silicon Valley. It also provided an opportunity to speak more freely than at public events, as participants agreed to share only conference information without mentioning anyone’s identity or job title. Scale’s summit could enhance its reputation as a leader in such discussions and place it at the nexus of the AI and defense worlds. Scale rented the hotel for the three-day event and footed the bill for most of the participants.
Scale agreed that Bloomberg could name some of the speakers who attended the event. They included Matt Knight, chief security officer at OpenAI; Andrej Karpathy, OpenAI researcher and former Tesla VP of AI; Craig Martell, Pentagon chief of digital and artificial intelligence; and Gen. James Rainey, commander of the US Army Futures Command. Also in attendance was Alexandr Wang, CEO and founder of Scale, who has been dubbed “Washington’s AI whisperer” and has been particularly vocal about highlighting the threats posed by China.
Wang said the focus of the conference was on rising geopolitical tensions — never mind that the event was held just as the leaders of the United States and China met to ease tensions. It was also boosted by the dramatic growth of functions and adoption of AI after the launch of OpenAI’s ChatGPT. “This past year has been absolutely amazing for anyone in the industry,” he told Bloomberg News. “More has happened in the last year than in the last 10 years.”
Although the focus was on examining the future opportunities and risks of the industry, there was no immediate change to the most famous artificial intelligence startup. Days after the conference, Sam Altman, CEO of OpenAI and the face of the AI industry, was ousted. At the time, there was no public indication from Karpathy or Knight that anything was wrong with the company.
One panel discussion focused on the AI race between the United States and China, and there have been signs of concern about China’s attitude toward Taiwan, the world’s semiconductor capital. Another talk offered a deep dive into the complexity of the chips and hardware needed to power AI, and a panelist explored the challenges of building data centers fast enough to meet rapidly growing customer demands. The third discussed how companies can secure their most powerful AI models. There was quite a bit of optimism the whole time. With avocado toast for breakfast and berry cocktails before dinner, the participants marveled at the innovation of artificial intelligence. They also questioned each other’s likelihood of AI progressing to a point where it can outperform humans, and debated whether the technology carries an existential risk. While these conversations can spark discussion on social media, one participant noted that in-person chats were more nuanced and reasonable.
Raquel Urtasun, CEO of autonomous driving startup Waab and a speaker at the event, said she will discuss with others how large AI models continue to scale and how AI will increasingly impact the physical world, including the rise of self-driving vehicles.
When attendees weren’t on panels, they enjoyed a bit of tech extravaganza, local style. They relaxed in a sprawling compound known as the “Exquisite Mountaineer.” They opted for custom cowboy hats and matching hat sets and dined on coffee-rubbed bison tenderloin served with wild mushrooms and blueberry juices.
On the second day, guests split into small groups and tried to hit a target that looked like a ferocious, life-sized bear, as well as strategically placed balloons. While checking out his teammate’s good hitting, one guest made a distinctly AI-themed joke to his group: “He didn’t need a lot of training data.”