Use AI assistants safely

An AI assistant can help with an outline, a draft or an explanation of a term. But the result may contain an error, an incomplete claim or an invented detail. Treat it as a draft to check, not as an answer that has been verified automatically.

Give it a clear goal, audience and format. Instead of “write something about the meeting,” try “suggest a brief list of points for a 20-minute team meeting.” Then adapt the output to fit your situation and voice.

Don’t enter passwords, identifying information, confidential company materials or other people’s personal stories into a public tool. Before using it, read about what data the service processes and what privacy options it offers.

For an important claim, ask for a source and open it separately. A link or a confident tone doesn’t prove that something is correct. If the assistant can’t substantiate a claim, verify it elsewhere or leave it out.

Use AI to support your own thinking, not to make decisions on someone’s behalf without their knowledge. Explain how the output helped if that matters for trust or workplace rules. You remain responsible for the content you send.

After editing the output, check names, dates, numbers and claims that could have consequences. Read it from the recipient’s point of view before sending. For work or public content, follow internal rules and be transparent about the sources used.

What’s a safe way to use an AI assistant?

Digital Support: A Practical Course

Continue the training

Enter your email address to continue the training.