A version of this story appeared in the daily Threat Status newsletter from The Washington Times. Click here to receive Threat Status delivered directly to your inbox each weekday.
Artificial intelligence specialists warned the Defense Department on Thursday that China has caught up in the AI arms race and has even pulled ahead of America in some areas.
The AI experts, who are working closely with the Pentagon, cautioned those attending a government-organized symposium on AI against overestimating America’s advantage in its high-tech competition with China.
Andrew Moore, a former Google executive advising U.S. Central Command on AI, said he’s seen China surpass America in several areas. He also said he respects the seriousness of the communist competitors’ work in the tech world.
“There are two application areas where they have outperformed hyperscalers in the United States. I’m not going to say where they are right now, what they are right now,” Mr. Moore said at the symposium, which was attended by defense and intelligence officers, tech companies, and other high-level stakeholders.
Hyperscalers are the Big Tech companies with computing power capable of meeting large demands from customers, such as Amazon Web Services and Mr. Moore’s former outfit, Google Cloud.
Large language models, or powerful algorithms, were a hot topic of conversation at the symposium, organized by the Defense Department’s Chief Digital and Artificial Intelligence Office. But Mr. Moore said he was more worried about the Chinese military investing in other powerful tools.
He said he would rank large language models fourth on his list of concerns. He said there are three other areas in which he expects China to push forward, though he did not identify them.
Alexandr Wang, CEO of Scale AI, which is partnering with the Pentagon on testing AI models, said the Chinese Communist Party is making rapid advancements to catch up on large language models and the resulting generative AI tools.
After missing the generative AI opportunity at the outset, China has aggressively pursued the technology in the last two years, according to Mr. Wang.
He pointed to a new AI model developed by a Chinese company run by Kai-Fu Lee, former president of Google China, as the best-in-class open-source model of its size.
“Still the best frontier models are being built in the United States by companies like Google and OpenAI but I think we need to treat them as very meaningfully caught up on the technology side,” Mr. Wang said at the symposium. “I don’t think they’ll surpass us on the technology, but they are caught up.”
Mr. Wang said the U.S. could fall behind if it fails to adequately invest in critical AI areas and if the U.S. does not build the right evaluation frameworks for responsible AI deployment.
Mr. Wang’s Scale AI just scored an undisclosed sum of taxpayer cash from the Pentagon to test and evaluate AI models. The entrepreneur’s team is planning to build benchmark tests for the department to apply to large language models working on massive data streams.
To win the AI race, Mr. Moore said the U.S. needs to out-compete its communist rivals on a wider range of tools than those exclusively involving large language models.
“We have got to get good at using advanced technology, whether it’s called AI or not, for automation, within the U.S. military and intelligence and with our Western allies,” he said. “That’s our goal. That’s China’s goal, to do the same for themselves.”
Mr. Moore is building a new AI startup in Pittsburgh named Lovelace AI, which intends to support missions in national security and disaster relief.
Editor’s note: The author of this article has no connection to the company Lovelace AI, which takes its name from Ada Lovelace, the 19th-century mathematician and trailblazer for the computing industry.
• Ryan Lovelace can be reached at rlovelace@washingtontimes.com.
Please read our comment policy before commenting.