What You Tell ChatGPT Be Used as Evidence in Court, Lawyers Warn

Kenyan legal experts have warned that conversations with artificial intelligence platforms may not be private, following a recent court ruling in the United States that highlights potential legal risks for users.

Lawyers caution that many users mistakenly believe AI chats are anonymous. Advocates Mary Audi and Fridah Muriithi of Muri Mwaniki Thige & Kageni Advocates explain that platforms often collect metadata such as IP addresses, device identifiers, and account details. 

Under Kenya’s Data Protection Act, 2019, this information can make users identifiable. Deleting a conversation does not ensure it is removed from servers or excluded from training systems. They emphasise that interacting with AI introduces a third party, and such exchanges do not carry the confidentiality of lawyer-client or doctor-patient relationships.

The storage of data across borders adds further complexity. Information entered into AI systems is frequently held in jurisdictions such as the United States or the European Union, limiting the direct authority of Kenyan regulators. 

Although access to such data in legal proceedings usually requires a warrant or court order, enforcement becomes more difficult once the data is stored abroad. Authorities may need to rely on Mutual Legal Assistance treaties, and the data may also fall under foreign legal processes.

Legal practitioners note that Kenyan courts already accept digital communications, including WhatsApp messages and emails, as evidence when they meet the requirements of the Evidence Act. AI-generated material, including prompts, responses, and chat records, may therefore also be admissible. 

However, the reliability of AI outputs remains uncertain. Since such systems can produce inaccurate or fabricated information, courts are likely to treat unverified outputs with caution.

User inputs may carry greater evidentiary significance. Prompts that indicate attempts to alter records or conceal information could be used as circumstantial evidence of intent. 

Courts are increasingly willing to order disclosure of digital records in civil and commercial cases, and AI-generated drafts may not be protected in the same way as formal legal advice. Lawyers warn that relying on AI tools in case preparation could create records that are subject to discovery.

Legal experts advise users to exercise care when using AI platforms. Under Kenyan law, entering personal information into such systems constitutes data processing. 

Once shared, this information may be stored overseas, accessed through legal channels, or disclosed under foreign law. Practitioners recommend avoiding the inclusion of identifying details, financial information, medical records, trade secrets, or confidential legal material. 

Users are advised to treat all AI interactions as communications with a third party rather than as private exchanges.