Courts are beginning to grapple with whether AI-generated material is legally privileged—and the answers are mixed. In this week's issue of NLJ, Stacie Bourton, Tom Whittaker & Beata Kolodziej of Burges Salmon examine US rulings showing how easily privilege can be lost
In one case, documents created using an AI tool were not protected because there was ‘no reasonable expectation of confidentiality’; in another, protection survived as disclosure was not to an adversary.
The analysis underscores a key risk: using public AI tools may amount to publishing information ‘to all the world’.
English courts are likely to take a similarly fact-specific approach, focusing on privacy expectations and tool settings. The warning is blunt—AI is ‘different’, and without careful governance, organisations may inadvertently waive privilege.
RELATED ARTICLES
