What is the purpose of ACL conferences?

2026-07-09 · Source: Ehud Reiter's Blog · Field: Science & Research — Research Methodology & Innovation, Artificial Intelligence & Machine Learning · Depth: Intermediate, medium

Summary

The ARR reviewing system for major NLP and ACL conferences faces immense strain from exponentially growing submission numbers, reaching 17K in the current cycle compared to 8K in an equivalent 2025 cycle. This pressure threatens the system's collapse, prompting a call for changes guided by a clear understanding of conference purpose. The author outlines four potential goals: facilitating researcher discussion, enhancing CVs, identifying high-quality science with precision, or identifying exciting science with recall. The current 3-month review process, involving three reviewers and a meta-reviewer, is struggling. The author suggests that most purposes, except for strict high-quality identification, imply a simpler, cheaper review process, potentially reducing human reviewers to two, eliminating author rebuttals, and integrating LLMs for tasks like finding hallucinated references.

Key takeaway

For research scientists evaluating conference submission strategies or considering volunteer reviewing, recognize that the ACL community is debating fundamental changes to its review process. Your efforts in reviewing may be misaligned if the conference's purpose is not clearly defined. Consider advocating for a simpler review system if your goal is broad dissemination or early-stage idea sharing, rather than strict quality filtering. This shift could lower prestige but improve system sustainability.

Key insights

The ARR reviewing system's sustainability requires clarifying ACL conference purpose to guide necessary process reforms.

Principles

• Conference purpose dictates review process design.
• High submission volume strains quality review.

Method

A simpler review process could involve two human reviews, limited LLM analysis, and no author rebuttal, potentially erasing main conference/"Findings" distinctions.

In practice

• Reduce human reviewers from three to two.
• Eliminate author rebuttals or shorten responses.

Topics

• ACL Conferences
• ARR Reviewing System
• NLP Research
• Peer Review
• Conference Management
• Publication Strategy

Best for: AI Scientist, Research Scientist, AI Student

Related on AIssential

• ACL ARR March 2026 Cycle [D] — The ACL ARR March 2026 cycle saw mixed review quality and significant procedural issues.
• A Simple Solution to Improve Broken Peer Review System at AI Conferences [R] — Splitting conference submissions and reviewers into independent halves can reduce reciprocal reviewing bias.
• "Taking Stock at FAccT": Using Participatory Design to Co-Create a Vision for the Fairness, Accountability and Transparency Community — Large-scale participatory design can effectively gather diverse community input for conference governance and strategic direction.
• ICLR’s Commitment to OpenReview — ICLR is doubling its financial support for OpenReview to promote open science and reproducible research.
• Should unis ditch group assignments? — Group assignments are vital for developing essential collaboration and co-regulated learning skills necessary for modern workplaces.
• Please follow the rules for ARR/ACL papers — Adherence to publication guidelines is crucial for research quality and reviewer efficiency.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Ehud Reiter's Blog.