Geneva’s parliament has drawn a firm line against unedited artificial intelligence in legislative drafting, warning elected officials that machine-generated texts risk democratic integrity and data security.
The Directive from the Sautier
Parliamentary leaders in Geneva are pushing back against an over-reliance on generative artificial intelligence tools among local lawmakers, Leman bleu reported.
The Bureau of the Grand Conseil decreed that any documents manifestement générés uniquement
by artificial intelligence must undergo substantial revision by their human authors before registration. Furthermore, the administration explicitly banned lawmakers from uploading confidential committee minutes and official documents into external artificial intelligence programs.
Divided Reactions Among Geneva Lawmakers
The intervention exposed sharp disagreements among political groups over the proper role of technology in public service. Ana Roch, head of the MCG parliamentary group, acknowledged that artificial intelligence has become part of daily routines for research and proofreading, but drew a hard line against full delegation. What we must not do is ask it to draft a text for the Grand Conseil,
she stated, according to Leman bleu.
Green Party deputy Julien Nicolet-dit-Félix adopted a far more critical stance, warning that machine-authored policy amounts to political deception. From the moment artificial intelligence drafts a text, replacing the elected person in terms of value, in terms of ideas, there is deception for the voters,
he said, as reported by Leman bleu. He also highlighted severe cybersecurity risks tied to corporate software platforms harvesting confidential committee data.
Meanwhile, PLR representative Jean-Pierre Pasquier offered a pragmatic counterpoint. He argued that artificial intelligence can serve as a useful starting point for parliamentary initiatives, provided the politician retains firm command of the subject matter and verifies all output.
Looking Toward Federal Infrastructure Models
Discussions in Geneva’s parliament have also turned toward how legislative bodies handle technology infrastructure. At the federal parliament in Bern, lawmakers already have access to a secure, custom-built artificial intelligence assistant designed specifically for elected officials—a technical model that Geneva could eventually adopt to mitigate private data exposure risks.
Worth a look
