International Courts Confront Litigation and Privilege Risks of Generative AI

Generative AI is reshaping litigation practice across common law jurisdictions, prompting international courts to reconsider how legal professional privilege risks, client confidentiality, and responsible AI use intersect.

Writing for Hall & Wilcox, Catie Moore and Lauren Separovich detail how Australia, the United Kingdom, and the United States are each developing distinct approaches, but common concerns are emerging around inadvertent waiver and the risks posed by public AI platforms.

Professional privilege safeguards confidential communications tied to legal advice or anticipated proceedings. Australia and the UK have both legal advice and litigation privilege. There are both attorney-client privilege and work product protection in the US. As generative AI tools become more accessible, particularly to self-represented litigants facing costly legal systems, questions have arisen over whether uploading privileged material into AI platforms compromises these protections.

Australian courts have issued a wave of practice notes following the Victorian Law Reform Commission’s report. They distinguish public from closed AI systems, though case law directly addressing privilege waiver remains limited. Decisions such as Helmold & Mariya and Mertz v. Mertz signal judicial concern over confidentiality breaches from AI use.

UK tribunals, including in UK v Secretary of State for the Home Department, have similarly cautioned that public AI tools may effectively place client information in the public domain.

US courts have taken a more fragmented, case-by-case approach, with rulings like United States v. Heppner finding waiver where public platforms were used without counsel supervision. Warner v. Gilbarco and related decisions protected AI-assisted work product prepared for litigation.

Practitioners advising on enterprise risk management and outside counsel spend should treat public AI tools as carrying meaningfully higher confidentiality and privilege exposure than closed, enterprise-grade systems. Given the international regulatory divergence, counsel managing legal operations across multiple legal systems must track jurisdiction-specific guidance closely. Disclosure obligations and professional conduct exposure remain live concerns wherever unverified AI outputs are filed with courts.