Five questions your AI-chat controls should answer

Why this matters now

In August 2026 the SRA published a warning notice on the misuse of AI. It says:

“Client information should only be entered into AI systems where appropriate contractual, technical and organisational safeguards are in place to protect confidentiality.”

The Upper Tribunal has drawn the same line between open and closed tools:

“Uploading confidential documents into an open-source AI tool, such as ChatGPT, is to place this information on the internet in the public domain, and thus to breach client confidentiality and waive legal privilege.”

“Closed source AI tools which do not place information in the public domain, such as Microsoft Copilot, are available for tasks such as summarising without these risks.”

Sources: SRA, Misuse of AI warning notice (17 August 2026); UK v Secretary of State for the Home Department [2026] UKUT 81 (IAC), headnote para 4 and para 21. General information, not legal advice.

What the checklist covers

1. 01Which AI-chat services and workflows are in scope?
2. 02Can confidential or privileged content be stopped before Send?
3. 03Does the control create another copy of the prompt?
4. 04Can you prove the control acted without reading employee chats?
5. 05How will the control be deployed, tested, and maintained?

Designed for a useful internal review

The checklist is educational rather than a compliance certificate. It helps IT, security, data-protection, risk, and legal teams compare policy with the controls and evidence available in practice.

Sovara Browser checks drafts locally before Send and records decision metadata rather than employee prompt content.