Pope Leo XIV proposes dual-track AI governance for global standards

Leo XIV, the chair of the European AI Forum, told a press conference in Brussels on Tuesday that policymakers must look beyond the rush to regulate generative AI and consider longer‑term governance structures. He urged governments, firms and civil‑society groups to coordinate on standards that can keep pace with rapid advances. The remarks came as the EU finalised its AI Act, the first continent‑wide set of rules for high‑risk AI systems, and as the United States prepared its own AI Bill of Rights.

The call for broader thinking reflects a growing frustration that, while most experts agree AI needs oversight, there is little consensus on what that oversight should look like. In the past year, high‑profile incidents – from deep‑fake political ads to biased hiring tools – have sparked public alarm. At the same time, tech firms argue that overly strict rules could stifle innovation and push development to jurisdictions with looser controls. The EU’s AI Act focuses on transparency, risk assessment and data quality, but critics say it leaves gaps in areas such as foundation models and cross‑border data flows. In the United States, lawmakers are split between a hard‑line approach that could limit research and a hands‑off stance that risks unchecked deployment. This split has left the global AI policy landscape fragmented, prompting calls for a shared, forward‑looking framework.

Leo XIV proposed a two‑track approach: immediate, enforceable rules for clearly hazardous applications, and a parallel, adaptive governance body that can update standards as technology evolves. He suggested the new body be composed of technologists, ethicists, regulators and consumer advocates, meeting quarterly to assess emerging risks. Industry groups welcomed the idea of a flexible mechanism but warned that any new bureaucracy must avoid duplication of existing agencies. The European Commission said it will study the proposal ahead of its next AI summit in November. If adopted, the model could become a template for other regions, offering a coordinated path forward in a field where disagreement has so far hampered decisive action.