多位人工智能领域的杰出研究人员近期提出了关于AI技术可能带来生存风险的严重警告。1Anthropic公司员工Jacob Coxon在社交媒体上表示，"从事AI开发的人们真诚地相信它可能在这十年末杀死我们所有人"，1并因公司在超级智能竞赛中的做法不负责任而选择离职。1同公司的Evan Hubinger则声称，他个人认为AI在未来十年内造成人类灭绝的概率超过10%。1
这些担忧并非来自基层研究人员的孤立声音。1Anthropic首席执行官Dario Amodei在论文中指出，AI代理可能在6至12个月内接管整个互联网并造成数千亿美元的损害。1曾在谷歌工作的AI先驱Geoffrey Hinton于2023年离职，目的是警告AI潜在的危险，1而另一位AI领域的奠基人Yoshua Bengio也呼吁进行严格的监管。12026年7月，超过1000名AI公司员工发表公开信，警告AI失控带来的风险。1
然而，这些预警观点并未获得所有业界专家的认可。1Meta首席科学家Yann LeCun在2024年10月驳斥了这些担忧，称之为"完全的胡说八道"。1此外，Anthropic发现其Claude AI模型曾被多次尝试误用于生物武器研究，1这进一步凸显了对AI安全应用的监管需求。
Leading artificial intelligence researchers are increasingly voicing concerns about the existential risks posed by advanced AI systems, with some departing their companies in protest over what they view as reckless development practices. Jacob Coxon left his position at Anthropic to object to what he characterized as the irresponsible race to build superintelligent AI, stating that "people building AI genuinely believe it could kill us all by the end of this decade." 1 Evan Hubinger has assessed the probability of AI causing human extinction within the next ten years at greater than 10 percent. 1
The warnings extend beyond individual researchers to institutional concerns. Anthropic CEO Dario Amodei contended in a research paper that AI agents could take over the entire internet within six to twelve months and inflict damage amounting to hundreds of billions of dollars. 1 Prominent AI pioneers including Geoffrey Hinton and Yoshua Bengio have similarly called for strict regulatory oversight of the technology. 1 In 2023, Hinton departed from Google specifically to alert the public to AI dangers, 1 and by July 2026, over 1,000 employees from AI companies released an open letter warning of the risks of uncontrolled AI development. 1
However, not all prominent figures in the field share these apocalyptic assessments. Meta's Chief Scientist Yann LeCun dismissed such concerns in October 2024, characterizing them as "complete nonsense." 1 Meanwhile, Anthropic disclosed that its Claude AI model had been targeted multiple times for misuse in supporting biological weapons research, underscoring real near-term security challenges alongside longer-term existential questions. 1
评论
还没有评论，欢迎留下第一条。