AI Ethics
What kind of impact would we like AI-driven systems to have on the world?
THE PROJECT IS COMPLETED
Systems driven by machine learning are already having a large impact on the world, and this impact is not always beneficial. The actively developing area of AI ethics is concerned with broader impacts of this technology, and ways to steer it in the desirable direction.
Rogers, A. Changing the World by Changing the Data? Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing (Volume 1: Long Papers). https://aclanthology.org/2021.acl-long.17
One of the factors in NLP making such fast progress since 2013 is the fact that resource development was also done quickly, without much consideration for who is represented in the data and how it is documented. The current turn to accountability and responsible AI requires that trend to be reversed.
Rogers, A., Baldwin, T., & Leins, K. (2021). 'Just What do You Think You’re Doing, Dave?' A Checklist for Responsible Data Use in NLP. Findings of the Association for Computational Linguistics: EMNLP 2021, 4821–4833. https://aclanthology.org/2021.findings-emnlp.414
SODAS researchers were part in the data governance working group of the BigScience project, which aimed to develop new ways to provide a responsible data governance infrastructure layer for training large language models.
The field of AI is increasingly coming under criticism for the demographic homogeneity of the people creating systems that have wide impact globally, and with little oversight. SODAS researchers were actively contributing to initiatives in making the publication process in NLP venues less biased towards the famous and well-resourced labs.
Panel: The Peer Review Process and Widening NLP. Bahar Mehmani (Elsevier), Anna Rogers (SODAS),Cecilia Superchi (Université de Paris). Widening NLP Workshop, co-located with EMNLP 2021. http://www.winlp.org/winlp-emnlp-2021/
How to review for ACL Rolling Review. Anna Rogers, Isabelle Augenstein. https://aclrollingreview.org/reviewertutorial
Anna Rogers contributed to the work on "Responsible NLP research checklist", adopted for submissions to ACL Rolling Review starting in December 2021. https://aclrollingreview.org/responsibleNLPresearch/
Funded by:
Copenhagen Center for Social Data Science (SODAS)
Full project name:
AI Ethics
Contact
Anna Rogers
Postdoc
SODAS
External members
| Name | Title | Phone | |
|---|---|---|---|
| Tim Baldwin | Melbourne laureate professor at the School of Computing and Information Systems, University of Melbourne | ||
| Kobi Leins | Senior Research Fellow, University of Melbourne | ||
| Isabelle Augenstein | Associate professer at the Department of Computer Science, UCPH | +45 93 56 59 19 |