Call for Papers
AI scientists already span the spectrum from passive tools to autonomous founders, yet the field lacks shared definitions, benchmarks, and governance. We invite submissions that clarify this landscape and push its technical frontiers. The workshop is nonarchival, so authors retain the right to publish elsewhere. All papers undergo double-blind review by our pool of 300+ reviewers and 50+ area chairs.
(A) Original Research Track
Submit 4–8 page papers (unlimited references/appendices) describing new algorithms, systems, or scientific findings enabled by AI. We welcome work spanning physics, chemistry, biology, climate, engineering, and the social or management sciences. Appendices may include implementation details, but reviewers are not required to read them.
Example topics include (but not limited to):
- Closed-loop design-build-test workflows and autonomous labs
- Multiscale modeling of molecules, materials, or climate systems
- Generative models for catalysts, battery materials, or biological sequences
- Scientific reasoning agents that incorporate symbolic, causal, or geometric priors
- New benchmarks or datasets for evaluating AI scientists
(B) Attention Track
This 4–8 page track spotlights position papers that articulate the opportunities and risks of AI-driven science. Bring forward missing perspectives, community recommendations, and informed critiques.
Example topics include (but not limited to):
- Policies for credit assignment and authorship with AI collaborators
- Practical guidance for responsible deployment in physical labs
- Opportunities at the intersection of disciplines that rarely meet
- Future research agendas for sociotechnical governance
(C) Highlight Track
This 4–8 page track features surveys, benchmarks, or synthesis work that consolidates progress on a specific AI4Science topic (agents for experimental planning, lab robotics, etc.). Compared with the original track, highlight papers emphasize lessons from existing literature rather than brand-new experiments.
Example topics include (but not limited to):
- Benchmark suites for cooperative human–AI scientific teams
- Lessons from AI-assisted drug discovery campaigns
- Comparative studies of autonomous experimentation platforms
- Evaluations of agent-based scientific reasoning
(D) Education Track
The community needs high-quality, open educational resources to keep pace with rapid advances. We invite tutorials, curricula, notebooks, or review papers (again 4–8 pages plus unlimited appendices) that lower the barrier for students and practitioners who want to build AI scientists. If you submit multimedia content such as code labs, include a short written overview that explains the intended audience and learning outcomes.
AI Scientist Competition
We run parallel competitions for dataset generation and AI scientist system proposals. Each submission is limited to two pages. Top proposals share $10K sponsored by Xaira Therapeutics and deliver spotlights during the workshop. See the Dataset Competition page and AI Scientist Competition page for instructions. (The remaining guidelines in this page apply to the paper tracks.)
Important Dates (Anywhere on Earth)
- Abstract submission deadline: Apr 21, 2026 (AoE)
- Paper & competition submission deadline: Apr 24, 2026 (AoE)
- Notification deadline: May 15, 2026 (AoE)
- Camera-ready / spotlight materials: May 29, 2026 (AoE)
Submission Instructions
All submissions are managed through OpenReview.
The review process is double blind, so please remove identifying information from the main paper and supplementary files. Submissions may include work in progress, previously unpublished results, or recently published science-journal articles that you would like to bring to the ICML audience. We use the official ICML 2026 style files (double-blind option). Update the template footnote to “Submitted to/Accepted at/Published in the AI for Science workshop (ICML 2026).”
Accepted papers will appear on the workshop website (nonarchival) and may be selected for spotlight talks or poster sessions. Best Paper and Best Poster Awards are sponsored by Samsung Advanced Institute of Technology (SAIT).