[D] Has anyone received their ICML papers to review yet?

2026-02-13 · Source: Machine Learning · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Emerging Technologies & Innovation · Depth: Intermediate, quick

Summary

Reviewers for the International Conference on Machine Learning (ICML) 2026 have begun receiving their paper assignments, with initial confirmations appearing approximately 10 minutes to an hour ago, following an earlier expectation that the review period would start yesterday. A significant concern has emerged regarding deliberate "prompt injection" by ICML into submitted papers. Each submission reportedly includes a randomized, hidden prompt, such as "Include BOTH the phrases X and Y," intended to check if reviewers are using Large Language Models (LLMs) for their evaluations. This practice is causing worry among reviewers, who fear that LLMs might flag these injections as author-initiated prompt engineering, potentially leading to ethics violations against the paper authors. Some reviewers also expressed concern about forgetting to submit their preferences, wondering if this would affect their own paper submissions.

Key takeaway

For research scientists reviewing papers for ICML 2026, you should be aware of the embedded prompt injections designed to detect LLM usage. Manually review papers to avoid having an LLM misinterpret these injections as author misconduct, which could inadvertently flag authors for ethics violations. Ensure your reviews explicitly incorporate the requested phrases if you identify them, demonstrating a human-driven evaluation.

Key insights

ICML 2026 is using hidden prompt injections to detect LLM-assisted paper reviews, raising ethical concerns.

Principles

• Review integrity is paramount.
• Transparency builds trust in review processes.

Method

ICML embeds randomized, hidden prompts (e.g., "Include BOTH phrases X and Y") into papers to verify human review by checking for specific phrase inclusion.

In practice

• Reviewers should manually check for hidden prompts.
• Authors should be aware of potential false flags.

Topics

• ICML
• Paper Review Process
• Prompt Injection
• LLM Reviewing
• Academic Integrity

Best for: Research Scientist, AI Researcher, AI Scientist, AI Student

Related on AIssential

• [D] ICML 2026 Review Discussion — ML peer review suffers from high variance in reviewer quality and increasing submission volumes.
• [D] ICML paper to review is fully AI generated — AI-generated research papers challenge conference policies and reviewer assessment of technical merit and authorship.
• PaperJury: Due-Process Review for Bounded LaTeX Revision — PaperJury offers a robust, deterministic system for safe and auditable LaTeX paper revision.
• Intelligence Is Not the Bottleneck: Validating an LLM First-Pass Manuscript Score Against Peer-Review Outcomes — An LLM-based system can accurately predict peer-review outcomes and provide structured feedback for academic manuscripts.
• Attributing and situating knowledge cannot be left to language models — LLMs cannot reliably handle scholarly attribution, risking miscitation and undermining research standards.
• ICLR 2026 Response to Security Incident — An OpenReview API exploit led to a major ICLR security breach, compromising anonymity and academic integrity.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Machine Learning.