Op-Ed: Brian Chau: How American policy must adapt to DeepSeek

Any American AI strategy must include accelerating American AI progress, Brian Chau writes in his latest op-ed.

DeepSeek, a leading Chinese AI company, went viral this week after releasing its R1 model. R1 is a competitor to OpenAI o1, demonstrating the Chinese firm’s ability to rival the best American companies in answering advanced scientific questions. In the word of the Economist, “China’s AI industry has almost caught up to America’s.”

DeepSeek completely changed how mainstream journalists, pundits, and policymakers think about AI. Over one week, it went from being a niche technical release to a national press cover story. The most important thing about DeepSeek might be what it signifies: a Chinese company has joined an exclusive club of top AI firms, formerly exclusive to American companies. Multiple independent validators have confirmed that DeepSeek’s R1 performance in math, science, and coding tests rivals top American models like OpenAI o1.

DeepSeek R1 is a “reasoning” model. It modifies a base model, DeepSeek V3, to answer more difficult questions using a sequence of outputs called “Chain of Thought.” Only three companies have a version of this sophisticated new AI technology: OpenAI, Google, and now DeepSeek. Other American AI players, such as Meta, xAI, and Anthropic, lag behind.

The biggest implication for American policy is clear: if American companies fail to continue improving, Chinese companies will surpass them. Regardless of the barriers we place on Chinese companies, Chinese AI will continue to improve. Consequently, any American AI strategy must include accelerating American AI progress.

To this end, President Donald Trump repealed the Biden AI Executive Order and signed another Executive Order to reverse the Biden administration’s agency rulemaking on AI. Using this authority, President Trump’s appointees can go further in promoting AI innovation by revoking partnerships with foreign regulators, reconstructing procurement guidelines to be focused on efficiency over ideology, and freezing grants to activists groups supporting state AI regulation.

While the Republican AI agenda appears to be deregulatory, state-level AI regulation threatens to weaken American AI companies. A bipartisan national framework focused on advancing American AI research can help ensure American companies maintain their lead.

DeepSeek has woken up American entrepreneurs and policymakers. Chinese AI is no longer a distant competitor. It is a near technological competitor. It is time for America to confront that reality. It is time for a reset, making advancing American research the number one priority of our AI policy.

Brian Chau is the executive director of Alliance for the Future.