Metsola warns EU lawmakers against substituting AI for debate
Planned European Parliament guidelines would require transparency, checks on AI-assisted work and safeguards for sensitive documents.
European Parliament President Roberta Metsola has cautioned lawmakers that using artificial intelligence does not transfer responsibility for their decisions, as the institution prepares new guidelines on the technology. Speaking at a parliamentary event, she said AI-generated amendments were no substitute for discussion among elected members, Politico Europe reported.
Metsola also distinguished AI-assisted legal research from the judgment needed for difficult decisions. She called for disclosure when the technology is used so voters can hold their representatives accountable.
Officials are concerned about a sharp increase in AI-written amendments, including some citing nonexistent laws. Parliament secretary-general Alessandro Chiocchetti described a case in which parliamentary lawyers spent two days establishing that legal advice an AI tool had given an MEP was incorrect.
Chiocchetti said lawmakers must retain responsibility for writing and scrutinizing legislation, although AI could assist staff in processing thousands of amendments. He said the administration would soon submit guidelines to parliamentary leaders covering disclosure, keeping sensitive documents out of public tools and checking work before circulation.
Staff already face disclosure requirements under rules introduced in 2024, but MEPs need not currently identify AI assistance with amendments, questions or speeches. Binding requirements for members may require amendments to Parliament’s procedural rules, a process involving its constitutional affairs committee.
Parliament is also developing the EPGenAI Hub on its internal systems for sensitive tasks. ChatGPT, Anthropic’s Claude and Meta’s Llama will remain available for less sensitive work. Built-in AI functions on MEPs’ tablets have already been disabled because of data protection concerns.
