Terms of Service Analysis  Terms of Service

OpenAI

Worth Noting
Analyzed by Clause Claw AI
Type: AI Services March 10, 2026
OpenAI's terms are standard for AI services but cap liability at just $100 and use your conversations for training unless you opt out.
Overview

OpenAI's terms of service govern ChatGPT, DALL-E, and their other consumer AI tools with provisions typical of the AI industry but some notable limitations. The agreement includes mandatory binding arbitration, meaning users cannot join class action lawsuits and must resolve disputes individually. OpenAI caps their liability at the greater of $100 or what you paid in the past 12 months, which could be inadequate for significant AI-related harm. By default, OpenAI uses your conversations and inputs to train their AI models, though users can opt out through their settings. The company disclaims responsibility for AI accuracy and specifically prohibits using AI output for important decisions about people like hiring, medical care, or legal matters. Users retain ownership of their content but grant OpenAI broad usage rights. Accounts created with corporate email addresses may be transferred to employer control. OpenAI can terminate accounts for policy violations using somewhat vague standards, potentially causing users to lose years of conversation history.

binding arbitrationlow liability capAI training on user dataAI accuracy disclaimersbroad content licensecorporate account controlvague termination policy
Key Concerns
Arbitration
All disputes must go through binding arbitration instead of courts → You can't join class action lawsuits and must resolve problems one-on-one with OpenAI
Liability Cap
Maximum liability is $100 or what you paid in 12 months → Even if their AI causes significant harm or data loss, they owe you at most $100
AI Training
They use your conversations to train AI models unless you opt out → Your private chats help improve their products, but you can disable this in settings
AI Accuracy
You can't rely on AI output for important decisions about people → Their AI might give wrong info about someone, but you're liable if you use it for hiring, medical, or legal decisions
Content License
They get broad rights to use your input for service operation → Standard license needed for AI to process your requests, but it's quite expansive
Account Control
Corporate email addresses can be transferred to employer control → Your boss could potentially access your ChatGPT history if you used a work email
Termination
They can suspend accounts for risk or policy violations → Vague standards could lead to account loss with years of conversation history
View original Terms of Service
Related Analyses
Clause Claw Terms of Service
Concerning
Analyze your own Terms of Service

Get AI-powered analysis of Terms of Service, EULAs, and privacy policies with the Clause Claw Chrome extension.

Download Clause Claw