The European Union told the UN General Assembly it wants the upcoming Global Dialogue on AI Governance to avoid splitting international rules on artificial intelligence into competing systems and instead bring coherence to existing initiatives, according to a statement delivered in New York.
The EU and its member states said they are preparing a written submission for the dialogue and will send it by the end of April.
The bloc called for the talks to be universal and inclusive, and for them to recognise existing international AI efforts in an “interoperable and compatible” way — meaning systems that can work together across countries.
It also urged the UN process to go beyond traditional plenary meetings and operate as a “multistakeholder” platform, with governments, industry, researchers and civil society taking part in interactive sessions focused on practical exchanges and concrete proposals.
The dialogue should help build UN members’ capacity and understanding of AI, including through knowledge-sharing that benefits all stakeholders, particularly those from developing countries.
Calls for child safety focus and changes to the draft agenda
The EU said the discussions should cover how AI can serve the public interest, including supporting the Global Digital Compact and the UN Sustainable Development Goals, as well as closing digital divides and promoting and protecting human rights.
It listed areas it said could be addressed in the dialogue, including international cooperation, transparency and accountability, safety and security, “information integrity”, accessibility and multilingualism, gender equality, environmental impact and ethical use of AI, with human oversight included in the approach.
Child safety and the protection of children’s rights in the use of AI should be explicitly addressed, it added, alongside a “human-centric” approach to the lifecycle of digital and emerging technologies.
The EU welcomed the overall structure proposed by the co-chairs for the first edition of the dialogue, but asked for more time to be allocated to multistakeholder sessions compared with plenary segments, saying the current plan includes more than one day of plenary discussions.
It also raised concerns about what it called duplicative sessions on interoperable AI governance on the first day, and suggested renaming a proposed thematic cluster to: “Respect for human rights and the responsible, accountable, transparent and human-centric approach to the lifecycle of AI.”
The bloc also asked for clarification on how the co-chairs plan to assign heads of delegation to each plenary and thematic session.
