Do Your Organisation’s Rules Cover the Use of AI?

For most governing bodies and professional bodies, the honest answer is no: rules written before generative AI became commonplace say nothing about how members, officials, panels or staff may use it, and that silence is starting to matter. AI does not need a governance industry of its own; it is a compliance question, and the place to answer it is in the rules the organisation already has.

Where does AI actually show up in a rulebook?

Rarely where people expect. The pressing questions are not about the organisation’s own technology strategy but about conduct and process: a coach using AI to draft athlete assessments, an official leaning on an AI tool for a decision, a member submitting AI-generated evidence or character references to a disciplinary panel, a panel secretary pasting confidential case papers into a public chatbot to summarise them, or a candidate using AI in an assessment for a professional qualification. Each of these engages rules the organisation already has, on conduct, integrity, confidentiality, evidence and fair process, and in each the question is whether those rules are clear enough to apply.

What should the rules address?

  • Conduct and integrity – whether, and when, using AI amounts to misconduct, for example in assessments, qualifications or competition
  • Evidence – how panels should treat material that may be AI-generated, and whether parties must disclose that they have used it
  • Confidentiality and data – what case or member information may never be entered into a public or third-party AI tool, consistent with data protection law
  • Decision-making – that decisions affecting members are taken by the people the rules empower, with any AI assistance checked and never substituted for their judgement
  • Staff and volunteer use – practical boundaries for the people who administer the rules, usually in a short policy made under existing powers

Does this need new rules, or amendments to existing ones?

Usually amendments, and fewer than people fear. Most conduct codes already prohibit dishonesty, most procedures already require decisions to be taken by the designated panel, and most confidentiality provisions already cover case papers; what they lack is wording that makes clear those provisions apply when AI is involved. A targeted amendment, made through the proper rule-making route and checked as part of a wider rules review, is generally better than a standalone AI policy that sits outside the rulebook and, in truth, binds no one.

Who carries the risk if nothing changes?

The organisation does. A panel that cannot say whether AI-assisted evidence is admissible, or a conduct code that cannot reach AI-enabled cheating, will find those questions argued over case by case, at cost, and with inconsistent results. Nor does the law wait for the rulebook to catch up: the UK Jurisdiction Taskforce’s Legal Statement on Liability for AI Harms (July 2026) confirms that ordinary principles of negligence already apply to the careless use of AI. We explore the wider point in AI is a compliance question, not a technology question.

If you’d like help bringing your rules up to date on AI, see our Regulatory Compliance advisory services.