In the absence of a legal framework governing the use of artificial intelligence (‘AI’) in alternative dispute resolution (‘ADR’), the Association of Arbitrators (Southern Africa) (‘AASA’) recently introduced AI Guidelines (‘the Guidelines’) in a bid to support the responsible integration of AI in arbitrations and adjudications. Specifically, the Guidelines aim to assist parties, arbitrators, and tribunals by outlining best practices that align procedural integrity with ethical awareness – and minimise risk.
Although not legally binding, the Guidelines offer practical direction. They encourage parties to agree in advance on whether AI may be used and whether the arbitrator or tribunal may issue directions concerning its use. This upfront agreement is imperative for transparency and consistency throughout the proceedings.
Undoubtedly, AI is increasingly used in ADR for conducting legal research, managing large volumes of documents, drafting legal text, and collating case chronologies. While these functions can enhance efficiency, AI use poses various risks, including data security threats and technical flaws. Indeed, arbitration is often preferred for its confidentiality, which could be undermined by improper AI use. In addition, AI tools are known to produce inaccurate outputs or ‘hallucinations’, including fictitious statutory provisions and case references, especially in legally complex situations.
In light of these concerns, it is essential that the deployment of AI be informed by several core principles articulated in the Guidelines.
First, accountability remains with the human arbitrators or adjudicators. Tribunals should maintain professional independence and rely on their own judgment, regardless of AI-assisted analysis.
Second, transparency around the use of AI is essential. Accordingly, parties should consider whether to disclose the use of AI to all participants.
Third, confidentiality is paramount, especially given that some public AI models may retain or learn from user inputs, which potentially compromises sensitive information.
Finally, fairness must be preserved, particularly as AI tools may reflect or amplify biases in their training data, which makes human oversight indispensable.
Ultimately, incorporating AI tools in the arbitration sphere should proceed in a manner that reinforces the core values of ADR, namely efficiency, confidentiality, procedural integrity, and fairness.
For expert legal guidance and advice, contact our alternative dispute resolution professionals at litigation@stbb.co.za.
This content is the property of STBB. We encourage the sharing of our content for informational purposes. However, if you wish to copy or reproduce our content on your own platform or website, please ensure that proper credit is given to STBB.
