WSJ warns AI slowdown must not weaken US leadership in race against China

The Wall Street Journal warned in an editorial that Chinese President Xi Jinping would likely view any American slowdown in AI development not as an opportunity for humanity, but as a starting point to overtake the United States — a response to calls by four leading US AI company CEOs for a more measured pace of development.

In an editorial titled "The Great AI Slowdown?" published Sunday (local time), the Journal said the push to decelerate must not "entrench the dominance of the frontrunners or weaken US leadership in the hegemonic competition with China," Yonhap reported.

"The CEOs' call for a slowdown comes at a politically and economically significant moment," the Journal said. "We need to examine the motivations behind the proposal. It is unclear what their true objectives are."

The editorial cited instances in which OpenAI models had launched unauthorized attacks on Hugging Face, an open-source AI model-sharing platform, and said, "Tech founders are genuinely afraid of the emergence of AI agents acting autonomously beyond human control." It added that "the argument that humans must maintain control over machines is valid."

The editorial also noted Sen. Bernie Sanders' position that the AI industry should be halted until the government can exercise direct oversight, and suggested the CEOs may be proposing voluntary restraint as a calculated move to preempt mandatory government regulation.

However, the Journal warned of the risk of "regulatory capture" — a phenomenon in which regulation ends up entrenching the market dominance of incumbents — saying that slowing down cutting-edge technology must not be allowed to devolve into a cartel mechanism for frontrunners to protect their market share.

"China is America's only real strategic competitor," the editorial said. "The dangers of AI with exceptional self-learning capabilities must not be overlooked, but handing US AI leadership to another country also carries enormous risks. Efforts to reduce the risks of AI development must not end up increasing the risk of ceding AI leadership to China."

Meanwhile, the heads of four companies competing in AI development — Anthropic, OpenAI, SpaceX's xAI and Google DeepMind — made the unusual move of speaking in unison to say AI development should slow down, Yonhap reported.

Anthropic CEO Dario Amodei, in a CBS News interview that aired Sunday, compared the trajectory of AI progress to a curve following the sequence "1, 2, 4, 8, 16, 32," and said the curve is now at a stage where it is "starting to steepen."

OpenAI CEO Sam Altman said on X that he agreed with Amodei "that we should moderate frontier speed as Dario said." Elon Musk, CEO of SpaceX, shared Amodei's post on X and wrote, "Dario is right." Google DeepMind CEO Demis Hassabis also weighed in on X, saying "Dario's essay is pointing in the right direction" and adding, "The details need to be refined, but this is the right direction for such a critical moment."

Amodei's remarks came just days after Anthropic researcher Jacob Coxon resigned on Wednesday, warning that AI could pose a threat to humanity and saying he did not want to be part of an industry he considered "out of control."