Using AI at Work Safely: A Privacy, Accuracy and Policy Checklist

Protect your data

• Do not paste passwords, API keys, personal data about customers, health records, unreleased financials or other confidential material into a tool unless your organisation has approved it.
• Read the provider's terms on data retention and training. Business or enterprise plans often differ from free consumer plans.
• When in doubt, remove names and identifiers or use made-up examples.

Verify before you rely

• Treat AI output as a draft from a fast but fallible assistant.
• Check every number, date, name, quote and link, especially anything that will be published or sent to a client.
• For legal, medical or financial matters, use the output only to prepare questions for a qualified professional.

Keep a human accountable

Decide who reviews AI-assisted work and who signs it off. Decisions that affect people, such as hiring, lending or discipline, should never be left to a model alone.

Be transparent

Tell clients, readers or colleagues when AI materially helped with a deliverable if they would reasonably expect to know, and follow any disclosure rules in your industry or platform.

Respect copyright and licences

Do not assume generated text or images are free of third-party rights. Check the tool's terms on commercial use, avoid prompts that ask for a living artist's style or protected characters, and keep records of how important assets were created.

Write a one-page policy

• List approved tools and the kinds of data allowed in each.
• List forbidden uses and the review steps required.
• Name a person to ask when a case is unclear.
• Review the policy every quarter, because tools and laws change quickly.