European Parliament President Roberta Metsola has issued a firm warning that artificial intelligence cannot substitute the critical human deliberation required in EU lawmaking. Speaking at a Parliament event, she underscored that elected members must retain full responsibility for legislative decisions despite growing reliance on AI tools.
- Metsola insists human debate is vital for EU legislation.
- New AI guidelines seek transparency and safeguards.
- Parliament developing its own secure AI platform.
What happened
European Parliament President Roberta Metsola addressed lawmakers at a Parliament event focused on AI's role in legislative work. She warned against overreliance on AI for drafting legal amendments, emphasizing that AI cannot replace the substantive debate among MEPs required for sound decision-making.
The Parliament is observing a notable increase in amendments generated or assisted by AI, some of which contained references to non-existent laws. This has caused concern over the accuracy and reliability of AI outputs, prompting calls for rules to ensure proper oversight.
Why it matters
Metsola's remarks underscore the fundamental responsibility of elected representatives to maintain control over lawmaking, highlighting that AI tools are aids, not substitutes, for human judgment. She stressed that any use of AI must be transparent so voters can hold lawmakers accountable for decisions.
The risks demonstrated by recent incidents — including erroneous legal advice from AI requiring extensive review by Parliament lawyers — reveal the limitations of current AI tools in understanding complex legislative frameworks. This raises concerns about unchecked AI influence on critical government decisions.
What to watch next
The European Parliament is set to introduce new guidelines aimed at governing AI use within its workflows. These rules will emphasize openness about AI assistance, prohibitions on inputting sensitive information into public AI services, and rigorous fact-checking before outputs are finalized or published.
To support secure AI use, the Parliament is developing its own AI platform called the EPGenAI Hub. This system will be hosted internally to handle sensitive tasks, while commercial AI services like ChatGPT and Anthropic’s Claude will remain available for less sensitive work. Future rule changes may require formal amendments to parliamentary procedures to mandate transparency for MEPs.