Recently, Anthropic CEO Dario Amodei issued a stark warning in a newly released policy paper: artificial intelligence could lead to "significant, persistent job losses," an impact that may not be short-term pain during technological maturation but rather an inherent feature of AI technology itself. Amodei stated that although industry efforts aim to mitigate the shock brought by technological change, there remains a "quite possible" scenario where AI causes structural, long-term employment loss. He believes this might be an inevitable result stemming from AI systems essentially replicating human cognitive capabilities on a broad scale. In other words, this is not a "bug" but a successful "feature" of the technology. Amodei has previously issued similar warnings, stating that AI could eliminate half of entry-level white-collar jobs within five years and push unemployment rates up to 10%-20%, urging businesses and policymakers to stop glossing over these risks. This latest policy article focuses more on how governments should respond if persistent unemployment arrives. His recommendations fall into two categories: mitigating harm and sharing benefits. Specific measures include establishing a more comprehensive statistical system for assessing AI's labor market impacts, providing wage insurance for workers forced to accept pay cuts due to technology, retaining tax incentives for enterprises, funding vocational skills training, and expanding mechanisms for matching employment opportunities. If AI indeed permanently reduces demand for human labor, governments may need to consider long-term income support, such as financing universal basic income (UBI) through taxes levied on relevant companies or increased capital gains taxes, while also exploring wealth redistribution methods like "universal capital accounts." Notably, recently several AI industry leaders, including Amodei and OpenAI CEO Sam Altman, have adjusted their rhetoric to emphasize productivity improvements and new economic opportunities rather than issuing unemployment warnings. Media outlets point out that this shift relates to these companies' upcoming public offerings. Acknowledging in his article that Anthropic helps companies "do more with existing teams," Amodei stressed that if AI dividends are so substantial, society must plan ahead for workers who cannot automatically share in those benefits.