Policymakers weigh U.S.-China AI competition after DeepSeek
Boosting federal R&D, sustaining export controls and creating a federal AI framework are a few suggestions experts made to help maintain the U.S. lead in AI.
Lawmakers are concerned about what Chinese startup DeepSeek's recent AI models mean for U.S. AI competition. Some worry that potential cuts to federal agencies responsible for tech R&D could affect the U.S.-China AI race.
Rep. Zoe Lofgren (D-Calif.) criticized cuts made by President Donald Trump's administration to federal agencies -- particularly those efforts led by the Department of Government Efficiency (DOGE) -- and expressed concern about further "substantial cuts" to institutions like the National Science Foundation (NSF) and NIST. Lofgren spoke during a House Research and Technology Subcommittee hearing on Tuesday.
"It's unfortunate that our nation's great scientific minds are now facing an environment of uncertainty, distrust and reduced funding," she said. "President Trump with his Musk DOGE hackers are wreaking havoc on our scientific enterprise, not just endangering our ability to lead in AI and so many critical fields, but actively sabotaging our leadership."
Rep. Jay Obernolte (R-Calif.) said during the hearing that maintaining U.S. leadership in AI development and deployment is not only an economic imperative, but also a national security requirement. He said the introduction of DeepSeek's AI models earlier this year represents a "concerning milestone," particularly as agentic AI capabilities evolve. Agentic AI, or models using AI agents that operate autonomously to complete tasks, is what some consider to be the next stage of generative AI.
The Trump administration has trumpeted several companies' planned AI investments in the U.S., including the Stargate AI collaboration with OpenAI, Oracle and SoftBank. The administration is also crafting an AI action plan to establish U.S. policy on AI.
For the U.S. to maintain its lead in AI as competitors like DeepSeek arise, some experts testifying during Tuesday's hearing said policymakers should consider boosting federal R&D in the technology, while others said the federal government should craft an AI policy that preempts a patchwork of state AI laws.
Maintaining U.S. AI lead
Obernolte said as U.S. citizens and companies increasingly share private and personal data with AI systems, the U.S. "can't allow DeepSeek and other Chinese Communist Party-controlled entities access to this information." U.S. lawmakers have considered banning DeepSeek, a reaction that Fu Cong, China's ambassador to the United Nations, earlier this year called "unwarranted."
"This risk will only grow as we enter the era of agentic AI, where AI systems will actively book our travel, manage our finances, analyze our health records and handle other sensitive personal affairs on our behalf," Obernolte said during the hearing.
It will be a big deal when we want to use AI to write our code, manage our energy systems, conduct scientific research and make military decisions.
Tim FistDirector of emerging technology policy, Institute for Progress
Indeed, Tim Fist, director of emerging technology policy at the Institute for Progress, testified that security investments should be the U.S.'s top priority, particularly when it comes to AI agents. Fist said research has shown that AI agents can have backdoors, allowing bad actors into enterprises or government agencies without detection. Supporting federal R&D through agencies like the NSF and NIST to ensure that AI data centers and semiconductors can't be easily sabotaged will be critical, he said.
"It will be a big deal when we want to use AI to write our code, manage our energy systems, conduct scientific research and make military decisions," Fist said.
Julia Stoyanovich, director of the Center for Responsible AI at New York University, also testified in support of greater federal R&D in both data governance and transparency for AI models. Establishing the National AI Research Resource and providing federal support for institutions like the NSF will be critical, she said.
"This demands bold federal action, sustained investment, policy direction and commitment to research," Stoyanovich said.
Adam Thierer, a senior fellow at the R Street Institute, said not only do DeepSeek's AI models pose a competitive threat to the U.S., but models introduced by Alibaba and Tencent also do. For the U.S. to compete, Thierer said, policymakers should embrace open source AI innovation and craft a national framework preempting state and local AI proposals. He said the U.S. should also ensure balanced copyright and data privacy policies, attract more AI talent, and ensure diverse energy markets for AI.
Meanwhile, Gregory Allen, director of the Wadhwani AI Center at the Center for Strategic and International Studies, said that to mitigate China's rise in AI, policymakers should consider strengthening federal agencies that administer export controls.
Ensuring that agencies have the staff and resources to "successfully implement those controls" will be critical in protecting the U.S. lead in AI against China, he said.
Makenzie Holland is a senior news writer covering big tech and federal regulation. Prior to joining Informa TechTarget, she was a general assignment reporter for the Wilmington StarNews and a crime and education reporter at the Wabash Plain Dealer.
