Is it OK to use AI at work? Questions to check first

Many people use AI at work quietly, unsure whether they are allowed to, or avoid it because they think it might count as cheating. There is no single answer for everyone. Whether it is OK depends on your employer's rules, the data involved, what your clients have agreed to, and how you use it. This page sets out the questions to check.

This is a general guide, not legal or employment advice. Rules differ by country, sector, employer and contract.

The short answer: using AI at work is often fine, and many employers encourage it, but check first. Look for a workplace AI policy, keep confidential and personal data out of unapproved tools, check any client or contract terms, be open with your manager about how you use it, and stay responsible for the result. If you cannot find the answer, ask.

Question 1: Does your workplace have an AI policy?

Many organizations now have rules on which AI tools are approved and what they may be used for. Check your intranet, handbook or IT guidance, or ask your manager. If there is a policy, it is the starting point. If there is none, that does not mean anything goes; it means you should ask. For what such policies usually cover, see AI policy for employees.

Question 2: What data are you putting in?

This is where most real problems start. Ask:

• Does the text include personal information about customers, colleagues or patients?
• Is it confidential: plans, prices, contracts, unreleased results, source code?
• Is the tool one your employer has approved for this kind of data?

If the answer to either of the first two is yes and the third is no, do not paste it in. See AI agents and personal data and, for the law on confidential business information, AI tools and trade secrets at work.

Question 3: Do client or contract terms say anything?

Some client contracts, confidentiality agreements or professional rules limit how work may be produced or where client material may go. If you work on client matters, check the terms, or ask whoever manages the relationship, before using AI on that work.

Question 4: Should you tell anyone?

• Your manager: being open about how you use AI avoids surprises and lets your team share what works. If you would be uncomfortable saying how a piece of work was produced, that is a signal to ask.
• Clients and customers: some expect to be told; your employer's policy or the client's terms may say. In the EU, AI systems that interact directly with people must generally be designed so people know they are dealing with AI (EU AI Act, Article 50).
• Published work: check your organization's rules on crediting or disclosing AI assistance. For ownership questions, see AI-generated work and copyright.

Question 5: Are you still doing the job?

Using AI to draft, summarize or brainstorm is usually no different from using a template or a spellchecker, as long as you check the result and stand behind it. AI can produce confident but wrong output, which NIST calls confabulation (NIST AI 600-1), so the responsibility for accuracy stays with you. Problems arise when output is passed on unchecked, or when AI is used for something your role is trusted to judge personally.

Question 6: Is it a decision about a person?

Using AI to help decide on hiring, performance, pay or discipline is a sensitive area with its own rules in some places, including the EU. Do not do this on your own initiative; it should only happen within your organization's approved processes. See AI at work and workers' rights and, for screening and interviews, AI in hiring.

Is it "cheating" to use AI at work?

At work, the test is usually not whether you used a tool but whether the result is accurate, allowed and yours to stand behind. Using AI openly, within the rules, with checking, is generally treated as using a tool. Using it secretly, with confidential data, or to pass off unchecked work is where trouble starts.

Once you know it is OK, see how to use AI at work for how to do it well, and checking AI output for the checking habit.