Abstract
AbstractMachine unlearning (MU) is often analyzed in terms of how it can facilitate the “right to be forgotten.” In this commentary, we show that MU can support the OECD’s five principles for trustworthy AI, which are influencing AI development and regulation worldwide. This makes it a promising tool to translate AI principles into practice. We also argue that the implementation of MU is not without ethical risks. To address these concerns and amplify the positive impact of MU, we offer policy recommendations across six categories to encourage the research and uptake of this potentially highly influential new technology.
Funder
Alma Mater Studiorum - Università di Bologna
Publisher
Springer Science and Business Media LLC
Reference60 articles.
1. Achille, A., Kearns, M., Klingenberg, C., & Soatto, S. (2023). AI model disgorgement: Methods and choices. Proceedings of the National Academy of Sciences., 121(18), e2307304121.
2. Albergotti, R. (2023). The secret history of Elon Musk, Sam Altman, and OpenAI. Semafor. March 24, 2023. https://www.semafor.com/article/03/24/2023/the-secret-history-of-elon-musk-sam-altman-and-openai
3. Bender, E. M., Gebru, T., McMillan-Major, A., & Shmitchell, S. (2021). On the dangers of stochastic parrots: Can language models be too big? In Proceedings of the 2021 (ACM conference on fairness, accountability, and transparency (FAccT) (pp. 610-623).
4. Blistein, J. (2023). Sarah Silverman leads class action copyright suit against ChatGPT. Rolling Stone (blog). https://www.rollingstone.com/culture/culture-news/sarah-silverman-copoyright-suit-chatgpt-open-ai-1234785472/
5. Bourtoule, L., Chandrasekaran, V., Choquette-Choo, C. A., Jia, H., Travers, A., Zhang, B., & Papernot, N. (2021). Machine unlearning. In 2021 IEEE symposium on security and privacy (S&P) (pp. 141-159). IEEE.