(2024/03/25 出刊)

  The European Parliament passed the world’s first Artificial Intelligence (AI) bill by a landslide vote on March 13th. The so-called “AI Act” contains a comprehensive set of rules to regulate AI systems, including powerful tools like OpenAI’s ChatGPT. The bill is expected to officially become law by May or June. By 2026, the law will be in full force.
  The far-reaching regulation passed with the support of 523 lawmakers in the European Parliament. Only 46 voted against. The AI Act takes a risk-based approach. That means the riskier the system (like medical devices and critical infrastructure), the tougher the requirements (like using high-quality data). There are also outright bans on the AI tools deemed to carry unacceptable threat. Violations of the AI Act could draw fines of up to 35 million euros (or roughly NT$1.2 billion), or 7% of a company’s global revenue. The EU says it’s worried that these powerful AI systems could “cause serious accidents or be misused for far-reaching cyberattacks.” They also fear that generative AI could spread “harmful biases” across many applications, affecting many people.

 3 月13日,歐洲議會以壓倒性票數通過全球第一個人工智慧法。這項所謂的「AI 法案」制定了一套完整規範AI系統(包括ChatGPT在內的強大A I 工具)的法律。該法案預計在5 月或 6 月正式成為法律;2026 年前全面上路。
 歐洲議會以523 票支持通過該法案;只有46 票反對。此AI 法案以風險高低為依據來管理。也就是說,風險愈高的系統(如醫療設備和關鍵基礎設施),受到的規範就愈嚴格(如使用高質量數據);若該AI 工具被認為造成難以接受的威脅,則會直接禁用。觸法最高可被罰款3500 萬歐元(約新臺幣12 億元),或是公司全球收入的7%。歐盟表示,他們擔心這些強大的AI系統可能「導致嚴重事故,或被濫用於影響深遠的網絡攻擊」。他們也擔憂,生成式AI可能會在許多應用程式中傳播「有害的偏見」,影響許多人。