Jeff Bezos-Backed Anthropic Pushes For Stronger US AI Safeguards, Saying Autocracies Like China, Russia Could Use AI To Expand Their Power: 'Some Very Dark Possbilities'

Anthropic CEO Dario Amodei has called for stronger U.S. AI safeguards, warning that autocratic regimes like China and Russia could use artificial intelligence to enhance their control and military capabilities.

What Happened: During an interview with Hard Fork that was released on Saturday, Amodei expressed growing concerns about China's rapid AI advancements, stating that the country is catching up to U.S. frontier AI labs faster than expected.

While he downplayed commercial competition, Anthropic CEO highlighted national security risks.

“We have these autocracies like China and Russia, and I’ve always worried, I’ve worried, maybe for a decade that AI could be an engine of autocracy,” Amodei said. "If enforcers are no longer human, that starts painting some very dark possibilities."

See Also: Google And China's Honor Deepen AI Partnership As Huawei Spin-Off Takes On Apple And Samsung With $10 Billion Investment

Amodei argued that AI development in China is not purely commercial but involves a government component, making it a national security concern. He pointed to U.S. export controls as a critical tool to limit China's access to cutting-edge AI chips.

"I was heartened to see that the Trump administration is considering tightening the export controls," he added, referring to ongoing discussions about restricting China's AI development capabilities.

Subscribe to the Benzinga Tech Trends newsletter to get all the latest tech developments delivered to your inbox.

Why It's Important: Amodei's comments highlight the growing geopolitical stakes in AI development. China's AI progress could give its government enhanced tools for surveillance, propaganda, and military applications, potentially altering the global balance of power.

His warning also comes amid increasing U.S.-China tensions over technology, with AI being a key battleground. If China achieves AI superiority, Amodei suggests, it could use the technology to strengthen authoritarian control and challenge democratic nations.

Previously, Google DeepMind CEO Demis Hassabis acknowledged that China’s AI model is “probably the best work” from the country, but downplayed its significance as a scientific breakthrough.

Nvidia CEO Jensen Huang also spoke about the increasing computational demands of next-generation AI models. He specifically named DeepSeek's R1, OpenAI's GPT-4, and xAI's Grok 3 as models that rely on reasoning and need significantly more computing power.

This is despite Nvidia's 17% stock decline in January, driven by fears that DeepSeek had optimized AI performance with lower costs.

Meanwhile, a Piper Sandler's research last week noted Meta Platforms Inc.'s AI models lagging behind DeepSeek.

Check out more of Benzinga’s Consumer Tech coverage by following this link.

Read Next:

• MrBeast Lost Millions On Amazon's ‘Beast Games'—Now He's Looking To Raise Hundreds Of Millions To Take His Brand To The Next Level

Disclaimer: This content was partially produced with the help of Benzinga Neuro and was reviewed and published by Benzinga editors.

Photo courtesy: Shutterstock