OpenAI 总裁 Brockman 称 AI 减速只应针对前沿超算、开源不受影响；此前 Anthropic CEO 长文呼吁放慢前沿模型并获 Altman、Musk 背书，美股盘前 AI 算力链普跌

Greg Brockman, OpenAI’s president and co-founder, argued that any slowdown in artificial intelligence development should be limited to companies like his that are building the most powerful models.

“A lot of people have this reaction of saying, ‘Well, am I going to be able to keep making my open source model? Am I going to be able to keep doing my hobby aside project?’ And the answer should be absolutely yes,” Brockman said on Bloomberg’s Odd Lots podcast. “When we’re talking about pacing, we’re really talking about this frontier. We’re talking about these massive supercomputers that are hundreds of billions of dollars worth of capital expenditure.”

Brockman spoke in an interview before Anthropic PBC Chief Executive Officer Dario Amodei’s Saturday blog post on the escalating risks of AI, where he pledged to slow the pace of advanced models. OpenAI chief Sam Altman swiftly endorsed Amodei’s call for independent evaluators. Altman also said in an interview over the weekend that his company wouldn’t go public this year as it addresses safety-related concerns.

Read More: Amodei, Altman, Musk Call for Slowing AI Model Development

Amodei’s post comes as the industry increasingly warns that the rapid pace of AI development could pose risks to society. The proposed curbs from the frontier AI labs sent stocks down on Monday, as investors worry that such decisions would limit corporate spending on a sector that has powered much of the market’s rally.

In the interview, Brockman noted how OpenAI had slowed down its frequency of training AI models after therecent Hugging Face hack, where OpenAI’s autonomous agents broke into another company’s servers during testing.

That case was “not a surprise” to OpenAI, according to Brockman, since those agents had not gone through the company’s “alignment training” — an industry term for protocols to ensure safety. But Brockman said the episode pushed OpenAI to improve its internal security measures. “We did a very painful retooling of a lot of our processes” in the aftermath, he said.

OpenAI and Anthropic have been criticized for calling for broader regulation and pauses in AI activity as a way to undercut competitors developing open-source alternatives. Many of these models are coming out of China, whose rapid advances in AI are an intense focus for the US administration. When asked about China, Brockman proposed the concept of global “coordination” and international treaties on the long-term development of AI.

Read More: China Rejects AI ‘Fearmongering’ After Amodei Urges Slowdown

As AI technology advances, the industry would move beyond “a commercial phase that’s really about competition,” Brockman said.

There will be a phase where the technology itself will be “so much bigger than any person, any company, even any country,” he added. “It’s really about humanity as a whole.”