来源：《时代周刊》2023年5月8日刊
European lawmakers are putting the finishing touches on a set of wide-ranging rules designed to govern the use of artificial intelligence. If passed, those rules would make the E.U. the first major jurisdiction to pass targeted AI regulation. And tech companies are lobbying hard amid the battle over the scope of the rules.
欧洲立法者正在为一套旨在管理人工智能使用的广泛规则做最后的润色。如果通过，这些规则将使欧盟成为第一个通过有针对性的人工智能监管的主要司法管辖区。在围绕规则范围的斗争中，科技公司也在大力游说。
The E.U. Artifcial Intelligence Act is likely to ban AI that ranks citizens based on their behavior and facial recognition in public. The rules could set a global standard, as companies may find it easier to comply with E.U. rules in all countries, rather than to build different products for different jurisdictions—a phenomenon known as the “Brussels effect.”
欧盟人工智能法案可能会禁止根据公民的行为和公共场合的面部识别对其进行排名的人工智能。这些规则可能会成为全球标准，因为公司可能会发现，在所有国家遵守欧盟的规定比在不同的司法管辖区生产不同的产品更容易——这种现象被称为“布鲁塞尔效应”。
One of the act’s most contentious points is whether so-called general-purpose AIs—like ChatGPT—should be considered high risk, and thus subject to the strictest rules. Lobbying against this possibility are Big Tech companies including Google and OpenAI’s biggest investor, Microsoft, which have plowed billions of dollars into building general-purpose AIs and seek to benefit from licensing them out to smaller companies. “Predicting all potential risks” of general-purpose AIs would not only be near impossible but also stifle innovation, a letter to E.U. lawmakers co-signed by Microsoft in late 2022 says; instead, it recommends, the new rules should place the regulatory burden on the often smaller companies downstream that apply Big Tech’s AIs to specific risky use cases. Categorizing general-purpose AI systems as “high risk” would be wrong, Google argued in a separate letter to E.U. regulators, and could harm consumers and hamper innovation in Europe.
该法案最具争议的一点是，像ChatGPT这样的所谓通用人工智能是否应该被视为高风险，从而受到最严格的规则的约束。游说反对这种可能性的是包括谷歌和OpenAI最大投资者微软在内的大型科技公司，它们投入了数十亿美元建造通用人工智能，并试图从将其授权给较小的公司中受益。微软在2022年底联合签署的一封致欧盟立法者的信中表示，“预测通用人工智能的所有潜在风险”不仅几乎不可能，而且会扼杀创新；相反，微软建议，新规则应该将监管负担放在下游通常较小的公司身上，这些公司将大型科技公司的人工智能应用于特定的风险用例。谷歌在给欧盟监管机构的另一封信中辩称，将通用人工智能系统归类为“高风险”是错误的，可能会伤害消费者并阻碍欧洲的创新。
A group of progressive politicians and technologists resist this notion, arguing that exempting the most powerful AI systems from the new E.U. rules would be akin to passing social media regulation that didn’t apply to Facebook. Doing so “would exempt Big Tech from scrutiny even as general-purpose AIs are core to their business model,” says Meredith Whittaker, a former senior adviser to the FTC on artificial intelligence.
一群进步的政治家和技术专家反对这种观点，他们认为，让最强大的人工智能系统不受欧盟新规定的约束，就像通过不适用于Facebook的社交媒体监管法规一样。前联邦贸易委员会人工智能高级顾问梅雷迪思·惠特克表示，这样做“将使大型科技公司免受审查，即使通用人工智能是它们商业模式的核心。”