Abstract

As Artificial Intelligence systems grow in complexity and scale, so do their environmental and societal impact. This workshop explores the urgent need for rethinking AI in terms of sustainability, with a focus on ecological and environmental issues. Presentations and discussions will rely on two pillars: Sustainability in AI, focused on reducing the ecological footprint of machine learning research and deployment, and AI for Sustainability, highlighting applications to global environmental challenges. With an emphasis on introspection of current practices, the workshop aims to foster responsible innovation, shared metrics, and encourage the community to design methods that are efficient, accessible, and viable with constrained resources.

Important Dates

Submission Deadline: ~~October 10th~~ October 17th, 2025, 23:59 AoE
Accept / Reject notification: October 31st, 2025

Submission link

Submissions will be managed through OpenReview:

https://openreview.net/group?id=EurIPS.cc/2025/Workshop/RethinkingAI
Note that newly created profiles on OpenReview without an institutional email will go through moderation process that can take up to two weeks.

Call for Voluntary Reviewers

We are currently seeking voluntary reviewers to assist in the evaluation of submissions. If you have relevant expertise and are willing to contribute your time and insights, please fill out our reviewer sign-up form: https://forms.gle/sVbzN9zs2UqyVc2n6
Reviewers must have an OpenReview account to participate. Your involvement will play a crucial role in ensuring the quality of the work presented at the workshop.

Call for Contributions

We invite submissions to the Rethinking AI workshop, which will bring together researchers, practitioners, and policymakers to discuss the dual challenges of applying AI for environmental preservation and making AI itself more environmentally responsible.

The workshop will feature two complementary tracks:

Track 1: Sustainability in AI

We aim to discuss the environmental footprint of AI itself. We encourage submissions that address:

Measuring and reporting the carbon, energy, and water footprint of AI models
Algorithmic and hardware efficiency (compression, pruning, quantization, frugal AI)
Lifecycle impacts of AI infrastructure (training, deployment, hardware production and disposal)
Socio-technical and governance perspectives on responsible AI development
Position papers reflecting on current research norms and proposing strategies for sustainable AI practices

Track 2: AI for Sustainability

We also welcome work that demonstrates how AI can contribute to tackling urgent environmental challenges. Submissions may include, but are not limited to, applications of machine learning for:

Climate science and modeling
Biodiversity monitoring and conservation
Agriculture and food systems
Renewable energy and grid optimization
Sustainable urban planning and mobility
Resource management and the circular economy

Submission Guidelines

We invite submissions of novel and unpublished work, as either short papers of at least 2 pages and up to 4 pages, or long papers of up to 9 pages. All page limits are excluding references and appendix. Submissions should follow NeurIPS formatting style, the checklist can be optionally left out.
Published papers can also be submitted in their original format, with the reviews included as supplementary material.
We welcome research papers, position papers, and case studies. Accepted contributions are non-archival (not published in proceedings) and will be presented as posters, with selected works invited for short spotlight talks. We expect at least one author to present at the workshop.