Beijing: A Chinese DeepSeek engineer has raised eyebrows by comparing the possibility of US artificial intelligence lab Anthropic controlling the world's most advanced AI to Adolf Hitler and Nazi Germany acquiring nuclear weapons before the Allies. Liu Shengyu, a machine-learning systems engineer for DeepSeek's V4.1 model, expressed his concerns in a blog post written in Chinese.
According to Anadolu Agency, Liu criticized Anthropic and OpenAI, arguing that they might not make frontier AI "open and affordable," a stance he claims contrasts with DeepSeek's commitment to open-source models. This statement comes amidst growing concerns from US AI executives and researchers about rapid AI development and the potential for future systems to operate beyond human control.
Over the weekend, Anthropic CEO Dario Amodei advocated for a slowdown in AI development and suggested increased oversight by proposing the inclusion of independent third-party evaluators within leading AI companies. He emphasized that any measures should not compromise commercial advantage.
The AI industry was further stirred last week when former Anthropic researcher Jacob Coxon resigned, criticizing leading AI firms for irresponsible behavior. Anthropic's alignment-science lead, Evan Hubinger, added to the discourse by expressing his belief that there is a more than 10% chance of AI causing human extinction within the next decade.
Liu's blog post reiterated his belief in making cutting-edge AI accessible to all, stating, "I believe cutting-edge intelligence should be made available to everyone in an open and affordable manner." He voiced his distrust of Anthropic and OpenAI, stating, "I do not trust Anthropic or OpenAI to do this; specifically, I dread the prospect of Anthropic controlling the most advanced AI or AGI. To put it dramatically, the consequences would be no less severe than Hitler acquiring atomic bomb technology before the Allies."
Liu's comments underline the broader concerns about AGI, which refers to AI systems capable of matching or surpassing human-level intelligence. He warned that as AI continues to evolve, social mobility could become increasingly challenging, stating, "you need the most powerful AI to break through class barriers."
In related developments, the Dongbuyaqiao Research Institute, an independent Chinese research platform, interpreted Amodei's remarks as potentially seeking greater control over industry regulations before they are fully established. Meanwhile, Chinese Foreign Ministry spokesperson Guo Jiakun emphasized that "fearmongering, confrontation and vicious competition" could impede efforts toward effective global AI governance.
A commentary by the state-run Global Times described Amodei's proposals as a "silent AI Cold War," suggesting they were designed to maintain the US technological lead while sidelining China from global AI governance. US President Donald Trump has dismissed calls to decelerate AI development, arguing that holding back could provide China with an advantage.