Chan Zuckerberg Initiative Awards Grant to OpenReview for AI-Enhanced Peer Review

Andrew McCallum

We are thrilled to announce that the Chan Zuckerberg Initiative (CZI) has awarded a new grant to support OpenReview’s next phase of innovation: Flexible Tools and AI Assistance for Efficient Scientific Peer Review Workflows. This project—led by Andrew McCallum, Distinguished Professor at UMass Amherst and President of the OpenReview Foundation—builds on OpenReview’s decade-long mission to make peer review more open, efficient, and fair through open-source technology and community collaboration.

## A Growing Challenge and a New Opportunity

As research output continues to expand exponentially, the burden on reviewers and editors has become overwhelming. Many leading conferences in computer science and related fields have seen submissions grow more than fifteen-fold over the past fifteen years, while the pool of qualified reviewers has not kept pace. OpenReview was founded in 2013 to bring openness and innovation to this process, and it now serves more than 1,500 publishing venues and 270,000 active monthly users, including nearly all the flagship conferences in artificial intelligence. With CZI’s support, we aim to take the next major step—leveraging AI tools and improved configurability to make peer review both scalable and sustainable.

## Two Key Directions for Innovation

The CZI grant will fund two complementary development efforts:
1. Expanding Access to OpenReview: We will design a new, intuitive configuration interface and a library of workflow “templates” for common reviewing styles—such as double-blind or open review, hierarchical review structures, and specialized ethics or topical tracks. These tools will make it dramatically easier for editors and program chairs to configure reviewing workflows independently, reducing the workload on OpenReview staff and enabling adoption across a wider range of scientific fields. Integrated AI-based configuration assistance will help venues customize workflows through conversational natural-language guidance.
2. Integrating Large Language Models (LLMs) into the Review Process: OpenReview will deploy LLM-based tools that assist authors, reviewers, and editors—helping authors spot issues before submission, reviewers refine their feedback, and editors summarize discussions and ensure decision consistency. Venues will be able to choose between locally hosted open-source models (for maximum data privacy) or commercial APIs.

## Safeguards and Responsible Deployment

OpenReview will emphasize data privacy, bias mitigation, and transparency in all AI-enabled tools. Our architecture will allow local hosting of models, ensuring data never leaves OpenReview’s control when privacy is paramount. We will publish documentation and best-practice guidelines for responsible use and continuously monitor the system for potential adverse effects.

## Sustaining Open Science for the Long Term

This support from the Chan Zuckerberg Initiative will help make OpenReview more accessible, scalable, and sustainable for the global research community. By combining our open infrastructure with AI assistance, we aim to strengthen peer review’s integrity and inclusivity—empowering editors and reviewers to focus on scientific insight rather than administrative overhead.

We thank CZI for believing in OpenReview’s mission and for investing in the future of open, fair, and efficient peer review.