围绕人工智能安全监管的行业争论日益升温，各方对如何推进AI开发的立场存在明显分歧。Anthropic首席执行官Dario Amodei撰写近4000字文章主张放缓AI开发并建立国际合作框架1，同时主张建立措施"在未来3-5年内足以放缓中国进展，显著扩大美国领先优势"1。OpenAI的Sam Altman和xAI的Elon Musk对Amodei的计划表示支持1，而Meta首席执行官Mark Zuckerberg表示"信任和对齐正在迅速成为最重要的能力"1，并延迟了Muse模型发布以专注于安全1。
美国政府和业界对AI监管的态度分化明显。Trump政府AI沙皇David Sacks表示AI监管应该由开发公司决定1，而美国众议院议长Mike Johnson则回应称"你们不会在10年内都死亡"1。主要AI公司正共同组建私人自律组织1，此举引发了关于"监管俘获"和竞争公平性的质疑。加拿大AI公司Cohere首席执行官Aidan Gomez指控大型AI实验室形成"卡特尔"1。
国际层面上，美国政府和中国似乎都无法停止AI军备竞赛2。中国外交部发言人郭家骏指责美国使用"恐吓论"破坏全球AI治理1。
The technology industry faces mounting tensions over artificial intelligence governance, with major companies and government officials taking sharply divergent stances on how to balance safety with competitive progress.1 Anthropic CEO Dario Amodei has authored an extensive article advocating for slowed AI development accompanied by robust safeguard deployment and an international cooperation framework.1 Meanwhile, companies including Meta and OpenAI have expressed support for market-driven self-regulation rather than government oversight, while the Trump administration's AI czar David Sacks has indicated that AI regulation should be determined by the developing companies themselves.1
The underlying tension reflects competing priorities regarding existential risk and competitive advantage.2 Tech leaders including Amodei, OpenAI's Sam Altman, and Elon Musk of xAI have voiced concerns about potential human risks from AI development, yet both the American government and China appear unable to halt an AI arms race reportedly involving $1 trillion.2 Meta CEO Mark Zuckerberg stated that "trust and alignment are rapidly becoming the most critical capabilities," and the company delayed its Muse model release to prioritize safety measures.1 However, House Speaker Mike Johnson dismissed existential fears, telling industry leaders "you will not all be dead within 10 years."1
International responses have grown contentious. China's Foreign Ministry spokesperson Guo Jiayun accused the United States of deploying "fearmongering rhetoric" to undermine global AI governance.1 Amodei acknowledged that proposed measures aim to be "sufficient to slow China's progress within 3–5 years, significantly expanding America's leading advantage,"1 while Cohere CEO Aidan Gomez has accused major AI laboratories of forming a "cartel" through their coordinated self-regulatory efforts.1
评论
还没有评论，欢迎留下第一条。