The EU Could Make ChatGPT and Other Ai Disclose Any Copyrighted Material Used in Training It

Under the draft AI Act, the developers of generative AI would have to disclose any copyrighted material which they had used to train their systems. This includes platforms such as ChatGPT and Midjourney. Under the new laws, AI tools would also be classed according to their risk level. While high-risk AI would not be banned it would be subject to greater transparency requirements helping the EU clamp down on biometric surveillance, misinformation and discriminatory language.