European Research Council Sets Dedicated AI Panel for 2028 Calls
The new PE12 route will evaluate foundational AI research separately from the ERC's broader computer-science panel when the next EU research framework begins.
Edited by Tyronne Panaino
The European Research Council announced on September 29, 2026 that it will establish a dedicated Fundamentals of AI evaluation panel for grant proposals. Named PE12, the panel is scheduled to begin with the 2028 ERC calls under the European Union's future research and innovation framework.
The change matters to researchers working on the foundations of artificial intelligence and to universities preparing future ERC applications. AI proposals are currently evaluated within the broader PE6 computer-science panel. A separate route should give foundational AI work a more explicit place in the review structure, although the announcement does not yet explain how proposals will be transferred, assigned or scored.
Why the ERC is creating PE12
The ERC says AI-related proposals now represent a significant share of submissions to PE6. Its response is to separate the field into a new thematic panel focused on fundamental and basic research rather than applications of AI inside another discipline.
The announced scope is broad. It includes machine learning, reasoning, intelligent and autonomous systems, computer vision, natural-language processing, generative and multimodal AI, and the mathematical and statistical foundations behind those areas. That definition puts model methods and core system research at the center of PE12 rather than tying the panel to one product category or industry.
What changes for applicants
The practical change begins with the 2028 calls, not the current funding round. Researchers whose main contribution advances AI foundations will have a named evaluation route. Work that applies AI to a specific scientific or technical domain can still be submitted to the relevant domain panel, according to the ERC's announcement.
That boundary will matter when a project combines method development with a substantial application. The press release states the high-level distinction but does not provide a decision tree for interdisciplinary proposals. Applicants and research offices will need the later call documents to see how the ERC defines panel fit, handles cross-panel review and treats work that spans foundational and applied objectives.
What the announcement does not settle
The ERC has confirmed the panel's name, intended scope and planned start. It has not yet published the panel membership, reviewer-selection process, scoring criteria, budget allocation or expected proposal volume. The page also links the change to the EU's future research and innovation framework, so the final 2028 call documentation remains the authoritative checkpoint for operational details.
This is therefore a structural funding-policy announcement, not evidence that any specific project will receive support or that total AI funding will increase. A dedicated panel may improve subject-matter alignment in review, but its effect cannot be measured until the ERC publishes the rules and begins evaluating applications.
Status
Confirmed. The ERC announced PE12 on its official site. Internal confidence is medium because the source is authoritative for the Council's own plan but no separate call text or implementation guide was available in the fetched evidence.
Sources
Update note: Last reviewed 2026-09-30. We will revise this post when the ERC publishes the 2028 call structure or panel guidance.
Sources
Drafted with AI assistance from source briefs; reviewed for citation completeness and label accuracy.