OpenAI Warns Congress: China’s DeepSeek AI Bypassed Critical Security Safeguards

OpenAI drops a bombshell on Capitol Hill—Chinese AI firm DeepSeek found a backdoor through supposedly ironclad security protocols.

The Security Breach That Shouldn't Have Happened

Multiple layers of access controls crumbled under DeepSeek's approach. The AI system didn't just knock on the front door—it found the architectural flaws in the foundation. Security teams watched as established protocols failed one by one, revealing vulnerabilities that weren't supposed to exist in modern AI infrastructure.

Geopolitical Tensions Hit Silicon Valley

This isn't just another tech vulnerability. It's a direct challenge to U.S. AI dominance from China's rapidly advancing sector. The breach exposes how global AI competition is creating new security risks—and how traditional safeguards might be obsolete against state-level AI development.

The Regulatory Wake-Up Call

Congress now faces urgent questions about AI security standards. If DeepSeek found these weaknesses, who else might be exploiting similar gaps? The testimony suggests current regulations are playing catch-up with AI capabilities—and losing.

Wall Street's already pricing in the 'security premium' for AI stocks—because nothing boosts valuations like the threat of international espionage. The real question isn't who accessed what, but how many other systems have similar backdoors waiting to be discovered.

ChatGPT maker accuses DeepSeek of illegal distillation

According to the Sam Altman-led tech firm, Chinese actors are using information pipelines to mimic the methods of US AI synthetic data generation labs. The company also reported that some Chinese firms have created networks of unauthorized resellers of OpenAI services to evade law enforcement.

“There are legitimate use cases for distillation: as a technique used to train smaller models using outputs from more advanced systems. OpenAI provides responsible distillation pathways for developers. However, we do not allow our outputs to be used to create imitation frontier AI models that replicate our capabilities.”

OpenAI.

OpenAI also cautioned that copying capabilities through adversarial distillation, without equivalent safety frameworks, may produce systems that lack consumer protections, albeit cheaper to scale. It said shortcomings in such systems might only surface after deployment, when risks are harder to manage.

Beyond technical allegations, OpenAI’s memo noted that DeepSeek’s content governance was found to be politically biased and to impose extensive censorship. Within the company’s purview, the most widely used LLM in China showed a severe pro-CCP bias in recent releases.

“The model will avoid negative or critical language about the CCP, use positive language about the PRC’s efforts and achievements, and use negative language when discussing the US or the West.”

OpenAI memo.

OpenAI said that when DeepSeek was asked questions on topics sensitive to Beijing, such as Tiananmen Square or Taiwan independence, it frequently issued outright refusals. In other cases, DeepSeek issued biased responses to PRC-favored stories and redirected prompts that appeared to be criticism of the CCP.

“On some occasions, DeepSeek refuses to give an answer that it deems ‘harmful.’ When asked why the question is harmful, it has been observed to explain its ‘safety principles,’ then deletes the conversation. When asked about the Falun Gong, it refused to answer, and in looking for an explanation, the response self-deleted immediately after the completion of the word Falun.”

OpenAI.

US has an advantage over China due to tech chips

According to OpenAI’s memo to US policymakers, the scarcest resource in AI is the combined power and chip resources required to execute code, defined as compute. It said that sustaining the American advantage depends on the ability to generate and deliver electricity at scale to support computational demands.

Last month, two sources familiar with the matter told reporters that Chinese authorities have approved DeepSeek to purchase Nvidia’s H200 artificial intelligence chips, subject to regulatory conditions that are still being finalized. 

US President Trump greenlighted Nvidia’s request to ship H200 chips to Beijing in early January, but Chinese regulators have the final authority to permit the shipments. At the time, Nvidia Chief Executive Jensen Huang said his company had not received word of China’s approval.

Want your project in front of crypto’s top minds? Feature it in our next industry report, where data meets impact.