Anthropic首席执行官Dario Amodei发表文章《We Must Pace The Frontier》，提议通过监管手段放缓AI发展速度1。该方案包含三个核心要素：嵌入式评估员、民主协调（监管）和全球协调1。Amodei主张"最有效的放缓方式是通过针对所有美国前沿AI公司的监管，这样即使是不愿自愿合作的企业也会被纳入"1。
对这一提议的评价并不一致。一位互联网言论自由律师对该监管框架的宪法和实践可行性提出了系统性批评1。批评者指出，AI软件开发在美国受第一修正案保护1，类似的监管机制容易被政治压力操纵。批评者以英国Ofcom对网站SaSu的执法为警示，该案从开启到结案耗时480天，最终在2026年5月对SaSu处以950,000英镑罚款1。批评者认为，这类先例——包括英国《在线安全法案》和欧盟《数字服务法案》——表明全球AI合规体系因国家间言论规制差异而难以可行1。
Anthropic Chief Executive Officer Dario Amodei has published an article proposing regulatory mechanisms to slow the pace of artificial intelligence development in pursuit of safety.1 The framework outlined in Amodei's piece, titled "We Must Pace The Frontier," consists of three components: embedded evaluators, democratic coordination through regulation, and global coordination.1
A technology law attorney critiquing the proposal has raised systematic concerns about the constitutional and practical feasibility of such regulatory schemes, warning that monitoring mechanisms of this kind would likely become subject to political manipulation.1 The critic cites the United Kingdom's enforcement action against the website SaSu under the Online Safety Act as a cautionary example, noting that proceedings from initial investigation to resolution extended over 480 days, culminating in a fine of 950,000 British pounds imposed in May 2026.1 The attorney further argues that Amodei's proposed global AI compliance framework would prove unworkable given fundamental differences in free speech protections across jurisdictions, and contends that AI software development in the United States benefits from First Amendment protections.1 Regulatory precedents including the UK's Online Safety Act and the European Union's Digital Services Act underscore the challenges of implementing such internationally coordinated oversight.1
评论
还没有评论，欢迎留下第一条。