Your team is already using AI. Your governance probably isn't there – yet.

AI governance for communications is the set of decisions a comms leader owns about how their team uses AI: when to disclose it, who signs off AI-assisted work, what must never be put into a tool, and who is accountable when something goes wrong. It is narrower and more practical than a generic corporate AI policy, because it is built around real communications tasks like drafting, research, monitoring and approvals.

Yes. Most comms teams adopted AI faster than they governed it, so the tools are already in the workflow while the rules are not written down. A short, specific written policy turns ad hoc judgement into something you can brief to your board and defend to a journalist, and the free toolkit on this page gives you a starting structure.

At a minimum it should cover disclosure (when you tell audiences AI was involved), accountability (who approves AI-assisted work), confidentiality and IP red lines (what must never go into a tool), and capability (how the team is trained). The Leadership AI Governance Toolkit organises these into four controls, each one usable in under thirty minutes.

Disclose when the use is material to how an audience would judge the content, for example AI-generated imagery, synthetic voice, or text presented as a named person's own words. Routine assistance such as research or first-draft support usually does not need a label, but the team should agree where the line sits in advance rather than case by case. The toolkit includes a disclosure decision guide for this.

Yes. The Leadership AI Governance Toolkit on this page is free, ten pages, and downloadable with no sign-up wall. It gives you a board-ready one-pager, a readiness check, a disclosure guide, an approval and accountability map, a claims and accuracy check, confidentiality red lines, and a thirty-day adoption path.

Shadow AI is the AI use inside your team that leadership cannot see: personal accounts, unapproved tools, work drafted somewhere off the record. It is usually a symptom of governance that restricts without offering a workable alternative, so the usage does not stop, it just moves out of sight. In communications it matters more than in most functions, because the output is public. The risk is not that someone used a tool you did not approve. It is that you find out about it from a journalist.

Accountability sits with the communications leader, the Head of Comms or Communications Director, not with the individual using the tool or with the IT function. AI does not approve its own work, so a named human signs off anything it touches. The toolkit accountability map helps you set this out for your real content types.