比尔·盖茨警告AI已越过危险阈值，呼吁加强监管与政策应对

比尔·盖茨在华盛顿州柯克兰接受《MIT Technology Review》采访时警告，人工智能在生物能力、网络能力、心理社会能力、就业市场破坏及失控风险等多个维度已越过危险阈值，而行业外对此缺乏足够关注 1。他特别指出，生物恐怖主义风险比自然大流行风险可怕且发生的可能性高出约50倍 1。

面对未来10至20年社会将经历的剧烈动荡期，盖茨提出多项政策构想以应对挑战，并认为距真正的丰裕至少还有十年 1。他呼吁对能够生成新分子的AI模型实施监控，并建议设立仅由人类从事的“人类保留岗位” 1。同时，他主张对AI使用征收“Token税”，将部分收入用于帮助因AI失业的人群 1。在监管机制上，盖茨认为不能指望AI行业自我监管，呼吁政府提升AI专业知识，并与中国合作制定监控标准 1。

In an interview with MIT Technology Review in Kirkland, Washington, Bill Gates warned that artificial intelligence has already surpassed critical danger thresholds across multiple dimensions, including bio-capabilities, cyber-capabilities, psychosocial capabilities, job-market destruction, and a lack of control 1. "We’ve crossed the threshold in terms of [AI’s] bio-capabilities, cyber-capabilities, psychosocial capabilities, job-market-destruction capabilities, and even the lack of control," Gates stated, emphasizing that these risks lack sufficient attention outside the industry 1. He specifically highlighted the severe threat of bioterrorism, noting, "I view bioterrorism risk, versus a natural pandemic, as about 50 times more scary, more likely than a natural pandemic risk" 1.

To mitigate these emerging threats, Gates argued that the AI industry cannot be relied upon to self-regulate, urging governments to enhance their AI expertise and collaborate with China to establish monitoring standards 1. He insisted that any AI model capable of generating novel molecules must be strictly overseen, asserting, "Any model that can make novel molecules should be monitored" 1. To address the economic impact, Gates proposed implementing a "token tax" on AI usage to fund assistance for those who lose their jobs due to AI, alongside the concept of "human-reserved jobs" that are exclusively kept for human workers 1. Looking ahead, he cautioned that society must navigate a "very tumultuous period" over the next 10 to 20 years, estimating that true abundance is still at least a decade away 1.