Dar es Salaam: Artificial intelligence has become an essential tool in conflict resolution. We assist with legal research, document review, translation, transcription, case management, and preparation of draft submissions. These technologies increase efficiency, reduce costs, and make arbitration more accessible. However, once AI begins to impact substantive aspects of the arbitration process, particularly the evaluation of evidence, legal reasoning, and drafting of arbitral awards, significant concerns regarding due process and enforcement will arise.
From a judicial practice perspective, the biggest concern is that AI should never replace independent human judgment. The legitimacy of arbitration rests on the parties’ belief that their dispute has been determined by an impartial arbitrator exercising independent reasoning. If arbitrators delegate important functions to AI, parties may legitimately question whether they received a fair hearing or whether the arbitral award truly reflects the arbitral tribunal’s own assessment of the evidence and applicable law.
Such concerns are directly related to fundamental procedural guarantees, such as the right to a hearing, transparency in decision-making, and the duty of courts to independently assess matters before them. If AI influences substantive reasoning without adequate human oversight, arbitral awards may be vulnerable to challenges based on the arbitral tribunal’s failure to make an independent determination or on the basis that the parties were denied a fair opportunity to address issues that affected the outcome.
We have already seen examples that demonstrate the risks associated with uncritical reliance on generative AI. In the Tanzania High Court in Mbeya, FAM v. ZAM, Marriage Appeal No. 212 of 2025, the lawyer cited six non-existent authorities generated by an AI tool. The High Court emphasized that advocates have a duty to courts, their clients, and the administration of justice to accurately represent the law, verify authorities, and use technology appropriately. The court further stated that misuse of AI could give rise to sanctions including adverse costs orders, quashing of proceedings, referral to professional disciplinary bodies, contempt proceedings and even criminal liability.
Similar concerns are emerging internationally. Courts in Kenya, South Africa, and Zimbabwe have all encountered fabricated citations generated by AI, confirming that technological innovation cannot replace professional competence and ethical responsibility. Although these cases arose in litigation, the lessons apply equally to arbitration. The arbitral award is
It could face challenges to recognition or enforcement if parties can show that the AI breached fundamental procedural guarantees, including the right to a hearing and the exercise of independent judgment by courts. Therefore, while AI remains a valuable assistant, it cannot become a decision maker.
While AI can significantly improve the efficiency of arbitration, certain functions require legal reasoning, judgment, and accountability that cannot be replicated by technology and should therefore remain exclusively within the purview of arbitrators.
These core adjudication functions include:
-Assess the credibility of witnesses and weigh evidence.
• Establish disputed facts.
interpretation and application of the law;
• Determine procedural fairness and admissibility of evidence.
• Exercise discretion throughout the process. and
• Draft and sign a final award with reasons that truly reflect the court’s own independent analysis.
AI may appropriately assist with administrative and preparatory tasks, such as organizing documents, translating materials, creating chronologies, summarizing evidence, conducting legal research, and preparing preliminary drafts, but must not make decisions that require judicial discretion or that determine the outcome of a dispute.
To protect due process and maintain the enforceability of arbitral awards, courts, arbitral institutions and tribunals must adopt clear safeguards.
• First, there should be institutional guidelines governing the responsible use of AI in arbitration, clearly identifying both permissible and prohibited uses.
Second, arbitrators and lawyers should independently verify the accuracy of all AI-generated legal authorities, fact summaries, and translations before relying on them.
• Third, if AI plays a significant role in the preparation of filings or court documents, appropriate disclosures should be considered where necessary to maintain transparency and maintain party confidence in the process. •
Fourth, all sensitive information input into AI systems must be adequately protected to protect privacy, confidentiality, and legal privilege. •
Finally, ongoing training must be provided to arbitrators, lawyers, and judges so that they understand both the capabilities and limitations of AI and can provide effective human supervision.
Also read: AGs value negotiation skills
In Tanzania, the judiciary has already taken significant steps towards responsible adoption of technology through e-filing, virtual hearings, digitization of court records, AI-assisted transcription and translation systems, and legal research tools such as Tanzlii. Recognizing that technological innovation must always be accompanied by ethical protection and judicial oversight, it has also developed training materials on artificial intelligence for judges and magistrates, including in partnership with UNESCO.
At the end of the day, we conclude that AI needs to remain a tool that supports human decision-making, rather than replacing it. The legitimacy of arbitration derives not from technology but from trust that disputes will be resolved by independent, fair and responsible arbitrators. Efficiency must not come at the expense of fairness. The continued evolution of AI requires safeguards that maintain due process, equality of parties, and integrity of the arbitration process.
