The AI Right to Unlearn: Balancing Privacy and Generative Systems
- M Oliveira e Carmo
- Jul 7
- 1 min read
Generative AI challenges the right to be forgotten. Large language models don’t store data as static records, they memorize patterns. Removing personal data isn’t as simple as deleting a row in a spreadsheet.
Enter machine unlearning: emerging methods let AI models “forget” specific data without full retraining. Techniques like Example-Tied Dropout (ETD) and Redirection for Erasing Memory (REM) are enabling certified, auditable erasure while respecting computational limits.
The right to unlearn is a new frontier in privacy, bridging human rights and AI governance. Organizations must implement layered safeguards, offering deletion where feasible and statistical guarantees where absolute erasure isn’t possible.
Privacy in the AI era isn’t just about protection, it’s about autonomy, dignity, and accountability.
Check https://iapp.org/news/a/the-ai-right-to-unlearn-reconciling-human-rights-with-generative-systems for more
Comments