Is ChatGPT safe to use? A plain English answer

Safe in what sense? That’s the question worth asking first, because there are at least three different safety concerns people have.

Is it safe for your data?

Be careful what you paste in.

OpenAI’s free version uses your conversations to train its models unless you turn this off. Go to Settings → Data Controls → turn off ‘Improve the model for everyone.’

More importantly: don’t paste confidential company information, client data, or personal data into a consumer AI tool. Not because ChatGPT will do something malicious — but because you’re putting sensitive information into a system your organisation doesn’t control.

The rule: use AI with your own words and your own ideas. Don’t use it as a repository for information that isn’t yours to share.

Is it safe from errors?

No. ChatGPT gets things wrong. Confidently.

It will sometimes state facts that aren’t accurate, cite sources that don’t exist, or produce figures that look right but aren’t.

Use it for first drafts and thinking, not as a source of verified truth. Always check anything factual before you rely on it.

Is it safe to use at work?

This depends on your company policy.

Many organisations now have AI use policies. If yours does, read it. If it doesn’t, the conservative approach: use AI to help you think and write, don’t paste in anything confidential, and don’t represent AI output as your own without reviewing it.

Microsoft Copilot is different — when deployed through your company’s Microsoft 365 tenancy, it operates within your organisation’s environment.

The honest summary

ChatGPT is safe for most professional tasks if you use it sensibly. The rule is simple: don’t paste in anything you wouldn’t put in an email to a stranger.

— Anna