Bipartisan bill seeks to ban Chinese AI from federal agencies

5 hours ago 1

WASHINGTON -- A bipartisan radical of lawmakers connected Wednesday vowed to support Chinese artificial quality systems retired of national agencies portion pledging to guarantee the U.S. volition prevail against China successful the planetary AI competition.

“We are successful a caller Cold War, and AI is the strategical exertion astatine the center,” Rep. John Moolenaar, the Republican seat of the House Select Committee connected China, said arsenic helium opened a proceeding connected the matter. “The aboriginal equilibrium of powerfulness whitethorn precise good beryllium determined by who leads successful AI.”

The proceeding connected Capitol Hill comes astir 5 months aft a Chinese exertion start-up called DeekSeek introduced an AI model that rivaled platforms from OpenAI and Google successful performance, but outgo lone a fraction to build. This raised concerns that China was catching up to U.S. contempt restrictions connected chips and different cardinal technologies utilized to make AI.

The ever-tighter contention is present a cardinal portion of the U.S.-China rivalry. And truthful overmuch is astatine involvement that the U.S. indispensable win, witnesses told the legislature panel.

The 2 countries are “in a semipermanent techno-security contention that volition find the signifier of the planetary governmental bid for the coming years,” said Thomas Mahnken, president and CEO of the Center for Strategic and Budgetary Assessments.

Jack Clark, co-founder and caput of argumentation astatine Anthropic, told the committee that AI has built-in values.

“I cognize that AI systems are a reflection of the societies that are built from. AI built successful democracies volition pb to amended exertion for each of humanity. AI built successful authoritarian nations will... beryllium inescapably intertwined and imbued with authoritarianism,” Clark said. “We indispensable instrumentality decisive enactment to guarantee America prevails.”

Earlier this year, Chris Lehane, OpenAI’s caput of planetary affairs, told reporters successful Paris that the U.S. and China were the lone 2 countries successful the satellite that could physique AI astatine scale. The competition, which helium described arsenic 1 betwixt antiauthoritarian AI and autocratic AI, is “very existent and precise serious,” and the stakes are “enormous,” helium said, for “the planetary rails of AI volition beryllium built by 1 of those 2 countries.”

The 2025 AI Index Report by Stanford University’s Human-Centered Artificial Intelligence halfway has the U.S. successful the pb successful producing apical AI models. But the study notes China is rapidly closing the show gap, reaching adjacent parity successful 2024 connected respective large benchmarks. It besides shows that China leads successful AI publications and patents.

At the hearing, Clark urged the lawmakers to support and fortify export controls of precocious chips to China. “This contention fundamentally runs connected compute,” helium said. The U.S. indispensable power the travel of almighty chips to China, Clark said, “or other you're giving them the tools they volition request to physique almighty AI to harm American interests.”

Mark Beall, Jr., president of authorities affairs astatine The AI Policy Network, said determination are “a fig of precise glaring gaps” successful the U.S. export controls that person allowed China to get controlled chips. Lawmakers earlier this twelvemonth introduced a measure to way specified chips to guarantee they would not beryllium diverted to the incorrect hands.

In different legislative step, Republican and Democratic lawmakers successful some the House and the Senate connected Wednesday introduced a measure to prohibition Chinese AI systems successful the national government.

“The U.S. indispensable gully a hard line: hostile AI systems person nary concern operating wrong our government,” Moolenaar said.

The No Adversarial AI Act, arsenic proposed, seeks to place AI systems developed by overseas adversaries and prohibition their usage successful the U.S. government, with exceptions for usage successful probe and antagonistic terrorism.

Read Entire Article