European lawmakers want Brussels to lead the push for a global AI deal, while calling for tougher safeguards against election manipulation and a new AI Observatory inside the Parliament.
MEPs on Parliament’s Constitutional Affairs Committee (AFCO) backed on Thursday a report that calls the EU to lead an international agreement on AI governance, modelled on nuclear non-proliferation. It includes stronger human oversight, new safeguards for elections and fundamental rights.
Its rapporteur, Emmanouil Kefalogiannis (EPP/GRC), said technological progress “cannot come at the expense of our democratic accountability and our fundamental values”, arguing that AI should remain a tool serving people and democracy. The initiative counted 21 votes in favour, three against and no abstentions.
A nuclear-style AI deal
The proposal compares AI with transformative technologies such as nuclear technology. It argues that its scale, speed and cross-border impact justify negotiations on a dedicated global convention or treaty.
Among the risks, one relates to AI influencing elections. The document points to Europe’s fragmented electoral framework as unsuited to dealing with AI misinformation and manipulation. Therefore, MEPs proposed stronger action against deepfakes, microtargeting and manipulation of voters.
Besides, MEPs supported that AI used by EU institutions should remain under human oversight. While they recognised that the technology itself can improve lawmaking by identifying inconsistencies, overlaps and risks of legal fragmentation, it should not replace human responsibility.
European Parliament AI Observatory
Lawmakers also propose a European Parliament AI Observatory. The goal is to provide technical and analytical support to parliamentary committees. The body should be budget-neutral, built from existing Parliament expertise rather than given regulatory or enforcement powers.
The report argues that AI now raises questions going beyond technical regulation, affecting public authority, democratic accountability and even the EU’s institutional balance.
Soon to face plenary vote
The report is expected to go before the full Parliament during the 19–22 October plenary session. It follows Parliament’s approval in March of an AI Convention, which placed fundamental rights, democracy and the rule of law at the centre of AI governance.
But the new AFCO report goes a step further. It calls for changes to the EU Charter of Fundamental Rights to address the AI era, including protection of personal identity and likeness, algorithmic transparency, digital self-determination and protection against algorithmic discrimination.
