Judiciary to introduce AI policy amid concerns of bot-generated filings

**Judiciary to introduce AI policy amid concerns of bot-generated filings**

Kenya's Judiciary is drafting a comprehensive artificial intelligence policy that would govern how technology is used within court operations, as registries across the country report a growing volume of filings suspected to have been generated by automated tools rather than qualified legal practitioners.

Chief Justice Martha Koome's office confirmed the policy is in its final stages of development, with rollout expected before the end of the financial year. The framework is intended to formalise AI use in case management systems, legal research databases, and back-office administrative processes, while establishing clear boundaries around what constitutes acceptable technology-assisted legal work.

The concern about bot-generated documents is particularly acute in the High Court's Constitutional and Human Rights Division, where public interest petitions have surged in recent years. Court registrars have flagged filings that contain coherent legal language but reveal factual errors, fabricated case citations, and inconsistencies that suggest minimal human oversight — hallmarks of large language model output used without verification.

The Law Society of Kenya has been brought into consultations and is expected to issue accompanying guidelines binding on advocates. LSK officials have warned that practitioners who submit AI-generated documents without disclosure or review risk disciplinary proceedings under the Advocates Act.

Kenya's judiciary has already invested in digital infrastructure through the integrated Court Case Management System introduced during the COVID-19 pandemic, which allowed e-filing and virtual hearings. The new AI policy builds on that foundation but introduces an accountability layer that current regulations do not address.

Legal experts welcome the move but caution that enforcement will be difficult without technical tools capable of reliably detecting AI-generated text, a challenge that courts worldwide are only beginning to confront.