ComReg has published Guidance on using AI in end-user dispute resolution involving consumers and businesses users of phone, broadband, and postal services, along with their providers that are parties to end-user disputes.
AI, when used appropriately, can be a valuable tool for end-user dispute resolution, helping parties to prepare submissions, organise information, improve clarity and overcome language or accessibility barriers. However, it may also produce inaccurate, misleading, incomplete, or fabricated content.
ComReg wants to ensure the timely processing and resolution of end-user disputes and requires transparency regarding use of AI in submissions. Inaccurate AI-generated content may delay the process or be treated as an abuse of process. Common issues include:
- References to unverifiable, non-existent or incomplete information
- Fabricated, inaccurate or misleading facts, references, or citations (“hallucinations”)
- Inaccurate citations of legislation, case law, or legal terminology
- Excessively long submissions that delay the resolution of disputes.
AI should not replace human judgement. Users of AI remain fully responsible for the accuracy of information submitted and must check facts, verify references, ensure submissions are accurate and relevant without solely relying on AI-generated content without human review.
In appropriate cases, ComReg may request information to be resubmitted in a different or shorter format and seek clarification about the use of AI in preparing submissions. These measures will help ensure a fair, prompt, efficient and transparent dispute resolution process while allowing parties to benefit from the use of AI.
