What New AI Policies Are Shaping ICML 2026 Submissions?

ICML 2026 Call for Papers

ICML 2026 paper submissions are now open! The deadline for full papers is January 28, 2026 (AoE). This year the committee emphasizes three key AI‑related points: large language models (LLMs) cannot be listed as authors, any prompt‑injection is prohibited (or the paper will be rejected), and AI reviewing is being expanded.

Authors of accepted papers may freely choose whether to attend the conference.

The original submission version of accepted papers will be made public.

A limit is placed on the number of reciprocal reviews each author can be assigned.

ICML 2026 explicitly permits the use of LLMs to assist writing or research, but authors bear full responsibility for any content that could be considered plagiarism or academic misconduct.

LLMs are not eligible to be listed as authors.

Any "prompt injection" is strictly forbidden and will lead to immediate rejection.

AI tools may be considered during review, but full delegation of reviewing to AI is not allowed.

Submission Format and Important Dates

All papers must be submitted as a single file: the main paper is limited to 8 pages, while references, impact statements, and appendices have no page limit. No separate supplementary‑material deadline exists; the camera‑ready version may add one extra page.

Key dates:

Submission site opens: January 8, 2026.

Abstract deadline: January 23, 2026 AoE (January 24, 2026 12:00 UTC).

Full‑paper deadline: January 28, 2026 AoE (January 29, 2026 12:00 UTC).

Submissions are made via OpenReview: https://openreview.net/group?id=ICML.cc/2026/Conference . Position papers have a separate OpenReview track.

Reciprocal Review Policy

Every submission must have at least one author who agrees to serve as an ICML reviewer. An author may be assigned as a reciprocal reviewer for at most two of their own submissions, unless an exception applies (e.g., the author already holds an AC/SAC role).

Authors with four or more submissions must agree to review for ICML; they may serve as reciprocal reviewers for up to two of their papers. Exceptions are granted for authors holding AC/SAC or other organizing roles.

Double‑Blind Review and Preprint Policy

All submissions must be anonymized and must not contain any information that could violate the double‑blind policy. Authors may post non‑anonymous versions on preprint servers such as arXiv, but must not promote the work as an ICML submission during the review period, nor cite the non‑anonymous version in the submitted manuscript.

Ethical and Conduct Guidelines

All forms of plagiarism are prohibited.

Prompt injection is forbidden.

Promoting the work as an ICML submission during review (e.g., in talks or on social media) is not allowed.

Any collusion between authors and reviewers, ACs, or SACs is strictly prohibited.

Submitting work that is substantially similar to already published or concurrently submitted papers violates the dual‑submission policy and may be rejected or removed from the proceedings.

Conference Details

The 43rd International Conference on Machine Learning (ICML 2026) will be held in Seoul, South Korea, from July 7–12, 2026.

Scope of Topics

General machine learning (active learning, clustering, online learning, ranking, supervised, semi‑supervised, self‑supervised learning, time‑series analysis, etc.)

Deep learning (architectures, generative models, theory, etc.)

Evaluation (methodology, meta‑research, reproducibility, human‑computer interaction, etc.)

Machine‑learning theory (statistical learning theory, bandits, game theory, decision theory, etc.)

ML systems (implementation, scalability, hardware, libraries, distributed methods, etc.)

Optimization (convex and non‑convex, matrix/tensor methods, stochastic, online, non‑smooth, composite, etc.)

Probabilistic methods (Bayesian methods, graphical models, Monte Carlo methods, etc.)

Reinforcement learning (decision and control, planning, hierarchical RL, robotics, etc.)

Trustworthy ML (reliability, causality, fairness, interpretability, privacy, robustness, security, etc.)

Application‑driven machine learning

All contributions must present original, rigorous research of significant interest to the machine‑learning community, supported by reproducible experiments or solid theoretical analysis, and situated within the broader scientific literature.

Reference links:

https://x.com/icmlconf/status/1986089104367308805

https://icml.cc/Conferences/2026/CallForPapers