{"id":77998,"date":"2023-04-28T12:29:32","date_gmt":"2023-04-28T09:29:32","guid":{"rendered":"https:\/\/forklog.com\/en\/?p=77998"},"modified":"2025-09-11T02:25:37","modified_gmt":"2025-09-10T23:25:37","slug":"eu-to-set-rules-for-generative-ai-reuters-reports","status":"publish","type":"post","link":"https:\/\/u1f987.com\/en\/eu-to-set-rules-for-generative-ai-reuters-reports\/","title":{"rendered":"EU to set rules for generative AI, Reuters reports"},"content":{"rendered":"

Generative AI developers will be required to disclose any copyrighted materials used to train their algorithms. The agreement was reached by the drafters of the European Union’s Artificial Intelligence Act, Reuters<\/a> reports.<\/p>\n

According to a source familiar with the matter, the provision was a late add-on drafted over the past two weeks. Some committee members initially proposed a complete ban on using licensed materials to train generative algorithms. However, they abandoned that in favour of a transparency requirement.<\/p>\n

\n

"Contrary to conservative desires to tighten surveillance and left-wing fantasies of over-regulation, the Parliament has found a solid compromise. It will enable proportionate regulation of AI, protect citizens’ rights, and also foster innovation and stimulate the economy," said MEP Svenja Hahn.<\/p>\n<\/blockquote>\n

Macquarie analyst Fred Havmeyer said the EU’s proposal was "tactful." Some, meanwhile, proposed an approach of "ban first, ask questions later."<\/p>\n

\n

"The EU is at the forefront of regulating artificial intelligence technologies," he told Reuters.<\/p>\n<\/blockquote>\n

The European Commission began drafting the AI regulation bill in 2021. Since then, generative AI algorithms such as ChatGPT have gained particular popularity.<\/p>\n

Members of the European Parliament agreed to move the document to the next stage, during which the final details of the bill will be worked out.<\/p>\n

Under the proposals, AI tools will be classified according to the expected level of risk: from minimal to unacceptable. Areas of concern may include biometric surveillance, disinformation dissemination, or discriminatory formulations.<\/p>\n

High-risk tools will not be banned. However, users of such tools will be required to adhere to the highest level of transparency.<\/p>\n

The latest amendment would require developers of generative systems such as ChatGPT or Midjourney<\/span> to disclose any copyrighted materials used to train their algorithms.<\/p>\n

Earlier reports indicated that the European Parliament was close to agreeing on proposals for the AI Act.<\/p>\n