Stored and breachable
Your conversations are saved on the provider and, like any database, can be exposed in a breach.
AI privacy · explainer
Are AI chatbotssafe?
For everyday use, mostly yes, but the real question isn't safety, it's privacy: what happens to everything you type. Here are the genuine risks of mainstream AI chatbots, and simple ways to shrink them.
The real risks
It's about your data, not danger.
These are the five privacy risks that apply to most AI chatbots by default.
Used for training
Many consumer chatbots may train on what you type unless you opt out.
Linked to your identity
Chats are tied to your account and email, not anonymous.
Disclosed on request
Content can be handed over under valid legal requests.
Sensitive data logged
Whatever you paste, passwords, health, secrets, may be retained.
How to reduce them
Share less, choose better tools.
Good habits
Practical steps that work on almost any chatbot.
- Don't paste secretsNo passwords, IDs, medical or financial data.
- Turn off trainingAnd use temporary or incognito modes.
- Delete old chatsReduce what is stored about you.
A tool that collects less
The strongest privacy is data that was never gathered.
- No account or emailNothing tying chats to your identity.
- History on your deviceNot in a company database.
- Zero retention, no trainingProviders discard your message after answering.
- No trackingNo cookies, no ad profiling.
FAQ
Frequently asked questions.
For everyday questions, generally yes. The real risks are about privacy, not danger: what happens to what you type. Chatbots can store your conversations, train on them, tie them to your identity, and expose them in a breach or legal request. The fix is to share less and choose privacy-first tools.
Five stand out: your chats are stored and can be breached; they may be used to train models; they are linked to your account and identity; they can be disclosed under legal requests; and sensitive data you paste can be logged. None of these is exotic, they are the default of most services.
Treat anything you type as if it could be stored and read by the provider. Avoid pasting passwords, financial details, medical records, or other people's private data. If you must work with sensitive material, use a tool that doesn't retain or train on it.
Turn off training in the settings, use temporary chats, avoid sensitive data, delete old conversations, and prefer AI that needs no account and stores nothing. Less collected data means less that can ever go wrong.
It is designed to minimise exposure: no account or email, history kept in your browser, GDPR-compliant providers with zero retention that don't train on your chats, and no tracking. There is simply far less data about you to lose.
Powerful AI,
minimal exposure.
No account, no tracking, nothing retained. NOLO is built so there is far less about you to ever go wrong.