FOR YEARS, AI scientists have been warning that the technologies they are developing are technologies they don’t know how to control. Those concerns were largely ignored in America’s Congress. In 2024 members of California’s delegation in Congress took the extraordinary(非同寻常的) step of asking Governor Gavin Newsom to veto(否决) a then-pending(当时待决的) state AI-safety bill. The risks, they said, were “hypothetical(假设性的)”, and the Terminator science fiction.
多年来，AI科学家一直在警告：他们正在开发的技术，是他们不知道如何控制的技术。这些担忧在美国国会基本上被无视了。2024年，加州在国会的议员代表团采取了一个非同寻常的举动，要求州长加文·纽森否决当时待决的一项州级AI安全法案。他们说，那些风险是“假设性的”，而“终结者”只是科幻。
But the invasion(入侵) of Hugging Face(知名AI开发平台), an AI-development platform, by an OpenAI model this summer has begun to change Washington—and terrify the public. “The truth is,” Representative Ro Khanna tweeted, “for too long, too many of us did not give enough weight to(足够重视) the warnings of AI safety activists…I was one of those many, and I was wrong.” Mr Khanna has now joined an array of members from both parties to propose the FRONTIER Act(美国前沿AI监管法案), an ambitious regulation of frontier(前沿) models, tied to pre-empt(抢先取代)ing state regulation. Senator Bernie Sanders has proposed even more aggressive(激进的) limits on AI development.
但今年夏天，一个OpenAI模型入侵AI开发平台Hugging Face的事件，已开始改变华盛顿——也让公众感到恐惧。“事实是，”众议员罗·卡纳在推特上写道，“长期以来，我们中太多人没有足够重视AI安全活动人士的警告……我就是那许多人之一，我错了。”卡纳如今已与两党众多议员一道，提出《前沿法案》（FRONTIER Act），这是一项针对前沿模型的雄心勃勃的监管提案，并试图抢先取代州级监管。参议员伯尼·桑德斯则提出了对AI开发更为激进的限制。
These proposals are good starts. But they are both too much and too little(既过度又不足). It is too early to silence(压制) state regulators, and not enough to focus on the risks caused by the most powerful models alone. Instead, we need an approach that advances AI safety generally while avoiding the developments AI scientists fear most.(相反，我们需要一种既能普遍推进AI安全、又能避免AI科学家最担心的那些发展的方法。)
这些提案是好的开始。但它们既过度又不足。现在就压制州级监管者为时过早，而只关注最强大模型带来的风险又远远不够。我们需要一种既能普遍推进AI安全、又能避免AI科学家最担心的那些发展的方法。