Scale AI CEO Says US Risks Losing AI 'Ammunition' Edge to China

The head of Scale AI, which helps companies like the designer of ChatGPT improve the data that feed their systems, said the US risks losing its edge to China in the quality of data used as “ammunition” for the powerful new tools.

The US is at a “critical moment” in its race for AI supremacy against China, Scale AI Chief Executive Officer Alexandr Wang said at summit for government officials organized by his company, a start-up he founded in 2016 after dropping out of MIT.

“We’re risking America falling behind,” Wang said, adding that China is investing more in AI, both in absolute terms and relative to its larger defense budget. In his presentation, Wang said China’s People’s Liberation Army invested at least $1.6 billion in the technology in 2020, compared to the US Defense Department’s roughly $1.3 billion.

Wang described artificial intelligence as an “unavoidable” technology that must be integrated into military operations to remain competitive with global adversaries. He compared AI to nuclear weapons as a technology that will reshape global diplomacy and power.

“I believe the next era of warfare and deterrence will be determined by AI,” Wang said. “The future of our world hangs in the balance.”

Senator Mike Rounds, a South Dakota Republican speaking at the same event, agreed that the US currently has the advantage in data collection and labeling, but warned that TikTok, owned by Beijing-based ByteDance Ltd. could give the Chinese Communist Party access to additional English-language samples to train AI systems.

Wang said the US must leverage the massive amounts of data collected by a vast network of military hardware — sensors, cameras, satellites — gathering information that can be harnessed to train AI. It will be ensuring the quality of those inputs, Wang said, that will “turn this hardware advantage into data advantage.”

Scale AI helps customers, from ChatGPT’s OpenAI Inc. and chipmaker Nvidia Corp. to Toyota Motor Corp. and the U.S. government, helps improve data quality to build better AI systems, according to the company’s website.