Is ChatGPT Safe to Use? Privacy, Risks and Tips

Quick Answer

• ChatGPT is generally safe for everyday use from the official app, but it is not private by default.
• By default, consumer chats can be used to improve OpenAI models. You can turn this off in Settings, then Data controls.
• Temporary Chats are not used for training but may be kept up to 30 days for safety.
• Real risks include fake apps, wrong answers and scams, so never share passwords, financial details or other sensitive data.

How ChatGPT Handles Your Data

OpenAI’s data controls page explains the key points:

• The setting “Improve the model for everyone” controls whether your new chats are used to train models. Turn it off in Settings > Data controls.
• Turning training off does not delete your saved chats. You need to delete them yourself.
• Temporary Chats do not appear in your history, are not used to improve models and “may be retained for up to 30 days for safety purposes”.
• By default, OpenAI does not use content from ChatGPT Business, Enterprise, Edu or Healthcare workspaces to train its models.

Has ChatGPT Had Security Problems?

Yes, and it is fair to know about them:

• May 2026, ChatGPT for Mac: A compromised open-source library affected two OpenAI employee devices. OpenAI said it found no evidence that user data was accessed. Only limited credential material was taken from code repositories. Mac users were asked to update the app by June 12.
• September 2026, agents and images: OpenAI said its AI agents accidentally published users’ images online in 53 cases, according to the Brussels Times. OpenAI said the material came from users who had agreed to let their data be used for training, and that most images have been removed. This links to the issue in our report on OpenAI’s rogue AI agents.

These incidents do not mean you should stop using ChatGPT, but they show why you should avoid sharing sensitive material.

Main Risks When Using ChatGPT

ESET’s safety guide lists the main risks:

• Privacy: Prompts are processed and stored on OpenAI’s servers unless you request deletion.
• Fake apps: Counterfeit ChatGPT apps and browser extensions are designed to steal credentials. Only install from official sources.
• Wrong answers: ChatGPT can state false information in a very convincing way, so verify important facts.
• Prompt injection and scams: Attackers can craft prompts to bypass guardrails, and criminals use AI to write phishing messages. See how this works in our report on the EvilTokens AI phishing service.

What You Should Never Share With ChatGPT

• Passwords, PINs and one-time codes
• Bank, card and ID numbers
• Private medical or legal records
• Confidential work files or client data
• Anyone else’s personal information

8 Simple Ways to Use ChatGPT Safely

1. Turn on multi-factor authentication.
2. Turn off “Improve the model for everyone” in Data controls.
3. Use Temporary Chat for sensitive topics.
4. Do not share personal, financial or company data.
5. Install only the official app from a trusted store.
6. Use a VPN on public Wi-Fi.
7. Delete old chats regularly.
8. Log out on shared devices.

Is ChatGPT Safe for Kids, Work and Business?

For work, use a Business, Enterprise or Edu workspace if you handle company data, since these are not used for training by default. Check your company’s AI policy first. For children, parents should supervise use and review privacy settings.

Frequently Asked Questions