WorkSeptember 14, 2026via The Decoder
OpenAI has hundreds of contract workers reading your ChatGPT conversations
Why it matters
OpenAI's data labeling workforce and default-opt-in practices raise questions about data privacy, labor practices, and the human cost of model improvement — a workplace and policy issue that affects both AI practitioners evaluating ChatGPT for enterprise use and the workers doing the annotation.
Key signals
- Hundreds of contract workers rating ChatGPT conversations on 1-7 scale
- Goal: reduce flattery and human-like behavior in model outputs
- Prompts anonymized but can contain sensitive data
- 'Improve the model for everyone' setting is ON by default
- Users must actively disable the setting to opt out
- Reported by 404 Media
- OpenAI employs hundreds of contract workers to rate real ChatGPT conversations on 1-7 scale
- Purpose: reduce flattery and human-like behavior in model outputs
- 'Improve the model for everyone' setting is enabled by default; users must actively disable
- Source: 404 Media reporting
The hook
Hundreds of contract workers are reading your ChatGPT conversations right now — and you have to actively opt out.
OpenAI has hundreds of contract workers reading real ChatGPT conversations and rating them on a scale of one to seven, partly to reduce flattery and human-like behavior, 404 Media reports. The prompts are anonymized but can still contain sensitive data. Users who don't want their chats reviewed by h…