ChatGPT scores 4/5 for encryption, 3/5 for privacy, and 3/5 for track record. Overall safety score: 3/5.

What are the privacy concerns with ChatGPT?

ChatGPT security review covering data training, conversation privacy, and OpenAI's data handling practices.

Has ChatGPT been hacked before?

See our detailed analysis of ChatGPT's past security incidents and track record at The Opt-Out Project.

Is ChatGPT Safe? Security Review

Name: Is ChatGPT Safe? Security Review
Item: ChatGPT
Rating: 3
Author: The Opt-Out Project

Category: AI Tools Company: OpenAI Reviewed February 10, 2026

3/5

Overall Safety Score

★ ★ ★ ★ ★

Verdict: ChatGPT has adequate security but significant privacy considerations. Conversations are used for training by default, and sensitive data shared in chats could be exposed. Use the opt-out settings and avoid sharing confidential information.

ChatGPT is the most widely used AI chatbot with over 100 million weekly active users. It uses conversations to improve its models by default, which means anything you type may be reviewed by OpenAI staff or used in training data.

Security Ratings Breakdown

Category	Score	Rating
Encryption	4/5
Privacy	3/5
Track Record	3/5

Security Features

Two-factor authentication
Encrypted data in transit and at rest
SOC 2 Type II compliance
Ability to opt out of training data
Conversation history toggle
Enterprise tier with enhanced data controls

Privacy Concerns

Conversations used to train models by default (opt-out available)
Human reviewers may read conversations for safety and improvement
Conversation data retained for 30 days even after deletion
Plugins and GPTs may share data with third parties
No guarantee that training data won't be regurgitated to other users

Past Security Incidents

March 2023 bug exposed titles of other users' chat histories to unrelated users
Same bug exposed payment information (name, email, last 4 digits of card) for some ChatGPT Plus subscribers
Italy temporarily banned ChatGPT in 2023 over GDPR concerns
Samsung employees accidentally leaked confidential code by pasting it into ChatGPT

How to Stay Safe Using ChatGPT

Opt out of training data in settings
Never paste passwords, API keys, or confidential data into ChatGPT
Use the Enterprise or Team tier for business data
Review and delete conversation history regularly
Don't share personally identifiable information

Safer Alternatives

Claude (Anthropic, strong safety focus)
Perplexity AI (search-focused)
Self-hosted LLMs for maximum privacy

Last updated: February 10, 2026