Straight answer
Is it safe to use ChatGPT on board papers?
Not on a personal account. Board papers hold price-sensitive, personal and confidential material, and a personal account puts them under terms the company has not agreed. The safe route is a tool the company has approved and configured, on terms it has read, with a clear board rule on what may go in.
How can the board tell whether it is under control?
Signs it is under control
- The company has a written rule on directors' use of AI tools with board material, and directors know it.
- There is an approved tool, set up by the company, that directors can use for summaries if they want to.
- The company secretary can say which tools have been checked and what their terms say about stored data.
Signs it is not
- Directors paste papers into whatever tool they use at home, and nobody has asked.
- The board portal allows downloads with no record, so nobody knows where papers end up.
- The AI policy covers employees but says nothing about directors.
What is the one question to ask?
Which AI tools may directors use with board papers, and on what terms?
A strong answer
A named, approved tool or a clear “none”, a summary of what its terms say about storing and reusing data, and a short rule in the board's own guidance.
A weak answer
“Directors are senior people, we trust their judgement.”
Why is this a real question?
Summarising a long board pack with an AI tool is tempting, and often useful. The question is where the text goes. Public AI tools differ in what they keep, how long they keep it, who can see it and whether it can be used to improve the service. Those terms vary between free and paid versions, between personal and business accounts, and they change over time. This page makes no claim about any one tool. The point is that you should know the answer for the tool you use before you paste anything into it.
Board papers often include unpublished results, deal terms, people matters and legal advice. Putting them into a service the company has not checked can breach confidentiality duties, data protection rules or market rules, whatever the tool’s terms say.
What might the board be getting wrong?
Many boards have approved an AI policy for staff and assumed it does not apply to them. Directors are often the heaviest readers of confidential material and the least supervised users of technology. If the board has never set a rule for itself, management cannot be blamed for not enforcing one. It is also unfair to ban everything and then expect directors to read four hundred pages on a phone.
What should the board ask for instead?
Ask the company secretary and the person responsible for information security to propose a short rule for directors: which tools may be used, on which accounts, and with what material. Ask them to check the terms of each tool and the company’s own policy, and to say plainly what they found. If an approved tool can be offered, offer it. Then put the rule in the board’s induction and review it each year, because the tools and their terms will change.