Diligence is the part that protects you. It is the responsibility you keep when you use AI: being honest about how the work was made, handling data carefully, and owning whatever goes out under your name.
What this means for your team
Three habits cover most of it:
- Be transparent. Be clear, where it matters, about what AI helped with. Different work needs different levels of disclosure, but pretending it was all you is a risk you do not need.
- Mind the data. Do not paste client names, personal data, or anything confidential into a tool you have not checked. Know what your tools do with what you put in.
- Stay accountable. You are responsible for the output, not the model. If it goes out under your name, the checking is your job.
It is worth writing a short, plain policy on disclosure and data so your team is not guessing case by case.
Try this
Write down your team's answer to two questions: what can we safely put into AI tools, and when do we say that AI helped? Even a few lines removes most of the day-to-day uncertainty.
Common questions about using AI responsibly
How do you use AI responsibly at work?
Responsibly comes down to owning the result rather than the model. Be honest, where it matters, about what AI helped with. Keep client names, personal data, and anything confidential out of tools you have not checked. And remember that if it goes out under your name, the checking is your job, not the AI's. It helps to write a short, plain policy on disclosure and data, so the team is not deciding case by case under pressure.
How do you use AI responsibly and ethically?
The ethical and safe side of AI use is what the fourth D, diligence, is about. It rests on three habits:
- Be transparent: be clear, where it matters, about what AI helped with, because people have a right to know when AI played a real part in work or decisions that affect them.
- Mind the data: know what your tools do with what you put in, and do not paste anything confidential into a service you have not checked.
- Stay accountable: you are responsible for the output, not the model, so you verify the facts and stand behind what goes out.
Different settings expect different levels of disclosure, but the responsibility to meet them sits with you, not the tool.
What data should you not put into AI tools?
Keep client names, personal data, and anything confidential out of any tool you have not checked. The point is to know what a service does with what you put in: who owns it, who can see it once it is shared, and whether your organisation allows it. Before sharing anything sensitive, check the tool's data protection policy first.
Do you have to tell people when you have used AI?
Where AI played a real part in the work, or in a decision that affects someone, they have a right to know. How much you say depends on the setting, so ask who needs to know, when to tell them, and how much detail makes sense. This is less about following a rule and more about keeping trust, so pretending the work was all you is a risk you do not need.