Is AI such a threat to the Kenyan legal profession?

Artificial Intelligence (AI) is rapidly transforming the delivery of legal services across the world. From legal research and document review to contract automation and predictive analytics, AI offers significant opportunities to improve efficiency, increase access to information, and reduce costs. Like all disruptive technologies, however, its adoption also raises fundamental questions about professional standards, accountability, confidentiality, and the administration of justice.

In Kenya, these questions have recently moved from the realm of theory into practical reality following judicial scrutiny of AI-generated court pleadings. The debate is therefore no longer whether AI can assist legal work.

Rather, the question is how AI should be regulated within the legal system and where the line should be drawn between technological assistance and the exercise of legal judgment.

A recent and widely publicised example underscores the importance of these concerns. In Nayan Mansukhlal Savla v Commission on Administrative Justice, the High Court, in a ruling delivered in April 2026, set aside a judgment after finding that the proceedings had been founded on AI-generated pleadings. The Court held that: “The court was misled into embracing pleadings which were machine-generated, and a judgment that flows from this illegal process cannot stand. The same must be set aside, which I hereby do. It does not matter how well the case is argued if the pleadings were illegally drafted.”

Litigation process

In reaching this conclusion, the Court emphasised that the use of artificial intelligence in legal proceedings cannot be introduced in an unregulated manner. It observed that any formal integration of AI into the litigation process ought to be guided by an appropriate legislative and procedural framework to safeguard the integrity of judicial proceedings and maintain public confidence in the administration of justice.

The significance of the decision should not be overstated. The Court did not hold that all uses of AI in legal practice are unlawful, nor did it establish a blanket prohibition on AI-assisted legal work. What the decision highlights is the lacuna between rapidly developing technology and procedural rules that were drafted before such technology existed.

The case nevertheless underscores an important principle: litigation is not merely an exercise in generating text. It requires legal judgment, ethical decision-making, procedural compliance, strategic analysis, and professional accountability. These are functions that AI systems are currently incapable of performing independently.

Kenyan law already regulates who may provide legal services and prepare certain legal documents. Section 34 (Unqualified Person not to Prepare Certain Documents or Instruments) of the Advocates Act Cap. 16 (the Act) restricts the preparation of specified legal instruments and legal work for gain by unqualified persons.

Although the provision was enacted long before the emergence of AI, the public policy considerations underpinning it remain highly relevant today. The legislation seeks to protect the public from inaccurate, incompetent, or irresponsible legal services and to uphold the integrity of the justice system.

Whether Section 34 of the Act applies directly to AI-generated pleadings remains an open legal question. AI is not a legal person and cannot itself be prosecuted, disciplined, or regulated in the same manner as a human practitioner. Nevertheless, the policy objectives underlying the provision support the broader proposition that legal drafting affecting the rights and obligations of litigants should remain subject to meaningful human oversight and accountability.

This distinction is important because AI systems can produce convincing but inaccurate legal content. Around the world, there have been numerous documented instances of AI platforms generating fictitious authorities, misquoting precedents, misstating statutory provisions, or providing legally unsound advice. While an experienced advocate may identify such errors before filing a document, an unrepresented litigant may not.

Administration of justice

Unlike an advocate, an AI platform owes no professional duty to the client, the court, or the administration of justice. It cannot be summoned to explain a pleading, sanctioned for misconduct, ordered to compensate a client for negligence, or subjected to disciplinary proceedings before the Advocates Disciplinary Tribunal. Professional accountability remains attached to the human user rather than the technology itself.

For this reason, Kenya would benefit from a clear statutory and regulatory framework governing the use of AI in legal practice. Parliament, the Judiciary, the Rules Committee, and the Law Society of Kenya should collaboratively consider reforms addressing the use of AI-generated court documents and legal services.

Such reforms might include:

1.Requiring advocates who use AI-assisted drafting tools to review, approve, and assume responsibility for all pleadings and legal documents filed in court.

2.Requiring disclosure where AI has substantially contributed to the drafting of court documents.

3.Establishing professional standards concerning verification of authorities, citations, and factual representations generated through AI systems.

4.Clarifying the extent to which self-represented litigants may rely on AI tools when preparing court documents.

5.Creating safeguards for confidentiality, privilege, data protection, and cybersecurity where client information is processed through AI platforms.

Importantly, the objective of the proposed reforms would be to ensure that AI remains a tool that assists lawyers and litigants rather than a substitute for the legal judgment and professional responsibility that the administration of justice requires.

Follow our WhatsApp channel for breaking news updates and more stories like this.

Pearlyne Omamo is Managing Partner at ADRA Advocates LLP