Anthropic的CEO Dario Amodei和政策主管Jack Clark再次就人工智能（AI）可能带来的大规模失业发出警告。他们认为，鉴于AI技术发展的指数级速度，未来五年内，AI有可能取代大量人类工作岗位，尤其是初级白领职位。Amodei预测，失业率可能因此攀升至10%至20%。Clark强调，面对AI可能带来的颠覆性影响，亟需制定相应的政策应对方案。两位高管均表示，有责任向公众坦诚AI发展的潜在风险，以便为未来的社会变革做好准备。

🤖 **AI可能大规模取代工作岗位**：Anthropic的CEO Dario Amodei和政策主管Jack Clark均对AI可能带来的失业问题表示担忧。Amodei预测，在未来五年内，AI有可能取代一半的初级白领工作，并将失业率推高至10%-20%的区间。这涵盖了法律、金融和咨询等多个领域。

💡 **指数级发展与政策应对**：Amodei指出，AI技术的指数级发展速度使得预测其确切影响变得困难，但其颠覆性足以引起警惕。Clark则强调，面对AI可能造成的“大规模颠覆”，必须制定出相应的政策来应对，以减轻其负面影响。

🗣️ **坦诚沟通的必要性**：两位高管认为，有责任向公众坦诚AI发展的潜在风险，以便社会能够提前认识到并为可能的变化做好准备。这种坦诚的沟通被视为应对AI挑战的重要一步。

Anthropic cofounders continue to warn that AI could soon replace a lot of human jobs.CEO Dario Amodei said he felt it was necessary to speak up due to the likelihood of disruption.Jack Clark, Anthropic's head of policy, said a policy response is needed.

Two cofounders of Anthropic haven't been shy with warnings about how AI could wipe out jobs.

CEO Dario Amodei and Jack Clark, who is head of policy at the AI research company, gave more insight into those warnings during an Axios event on Wednesday, saying they felt compelled to be honest.

"As with most things, when an exponential is moving very quickly, you can't be sure," Amodei said at the Axios AI+ DC Summit. "I think it is likely enough to happen that we felt there was a need to warn the world about it and to speak honestly."

Clark said a policy response was needed due to "the scale of disruption we expect in the next five years."

The Anthropic leaders are among the tech and business leaders who have warned that AI will replace human jobs, especially entry-level office jobs.

Amodei told Axios in May that he thinks AI could eliminate half of all entry-level white-collar jobs within the next five years. He also predicted that the unemployment rate could jump to somewhere between 10% and 20% in that timeframe. He told the BBC earlier this month that the AI-induced entry-level job losses could occur in fields like law, finance, and consulting.

Amodei was asked about his (p)doom number, a term used to signal how likely it is that AI will destroy humanity. He also told Axios on Wednesday that he thinks there's a 25% chance the AI will go "really, really badly," and that there's a "75% chance that things go really, really well."