By Samy Bengio, Sasha Luccioni, Inioluwa Deborah Raji, chairs of the Code of Ethics committee
The increasing real-world impact of ML research and applications increases both the likelihood of meaningful social benefit as well as the attendant risk of harm. This harm can often be minimized by practical improvements to research practice – including improved transparency, better documentation and the more purposeful dissemination of ML models and datasets.
This week, NeurIPS adopted a Code of Ethics to help guide our community towards higher standards of ethical conduct, including research ethics and the broader societal and environmental impacts of our work. The code is the result of a multi-year effort to establish guidelines and norms for submissions to NeurIPS. As such, it outlines conference expectations about the ethical practices that must be adopted by submitting authors, members of the program, and organizing committees.
The Code of Ethics complements the NeurIPS Code of Conduct, which focuses on professional conduct and research integrity issues, including plagiarism, fraud and reproducibility concerns. It also informs the NeurIPS Paper Checklist and Code and Data Submission Guidelines, which outline more concrete communication requirements around ML artifacts submitted to NeurIPS.
The Code of Ethics as it exists today is an initial version that will evolve over time, as new approaches and applications of ML are developed. We will continue gathering feedback from the community and will use it to update the code over the coming years, to better reflect the norms and values of the NeurIPS community.
Thank you to those who contributed towards the code of ethics: Alina Beygelzimer, Corinna Cortes, Kate Crawford, Thomas G. Dietterich, Barbara Engelhardt, Jeanne Fromer, Iason Gabriel, William Isaac, Amanda Levendowski, Jason Millar, Shakir Mohamed, Nyalleng Moorosi, Alice Oh, Marc’Aurelio Ranzato