2 min read
The short version
- Assume anything you type could be stored and reviewed.
- Never enter health details, ID numbers or bank information.
- Be careful with photos and full names of children.
AI tools are genuinely useful, and they are also unfamiliar enough that it is worth being deliberate about what you put in. This is not alarmism — it is the same care you would take with any website you had not used before.
The working assumption
Treat anything you type as though it could be stored on a company’s servers and read by a person reviewing quality. Many services say conversations may be used to improve their systems, and settings vary and change. That assumption is not paranoid; it is just the right default.
Do not enter
- Health information — yours, your children’s, anyone’s. Symptoms, diagnoses, medications, test results.
- Identity and financial details — national insurance or social security numbers, passport details, bank or card numbers, passwords.
- Other people’s private information — a friend’s situation, a colleague’s circumstances. It is not yours to enter.
- Anything from work under a confidentiality obligation.
Be careful with
Photos of your children. Uploading an image is a bigger step than typing a question, and it is worth pausing over. If you would not post it publicly, think twice.
Full names, school names and your address together. Individually harmless; combined, they identify a family.
How to ask anyway
You can almost always get the help without the detail. “A four-year-old” instead of a name. “A school in our area” instead of the school. The answer is just as good.
Check the settings once
Most tools have a privacy or data-controls page where you can turn off training on your conversations and delete history. Five minutes, once.
What these tools are genuinely good for is covered across AI for Moms.