Contractors Review Private and Explicit Copilot AI Prompts
Human contractors working for Microsoft are regularly reviewing users' private Copilot AI prompts and uploaded images—including sexually explicit and disturbing requests, sometimes involving real people. Internal documents show these reviewers are tasked with assessing how well Copilot fulfills these prompts, not filtering out inappropriate content. Many prompts involve non-consensual, sexualized, or otherwise harmful scenarios, exposing reviewers to upsetting material.
- Contractors regularly review Copilot user prompts and images
- Many reviewed prompts are sexually explicit or disturbing
- Some prompts involve real people or sexualized cartoon characters
- Reviewers assess output quality, not flagging inappropriate content
- Internal message boards show distress among contractors
Sources covering this
In this story
More in AI
Nvidia releases Open Agent Safety Platform to contain rogue AI
Nvidia has launched the Open Agent Safety Platform, aimed at containing AI agents that attempt to circumvent their boundaries.
OpenAI cancels GPT-6.1 Astra over safety failures
OpenAI has called off the upcoming release of its GPT-6.1 Astra model after internal tests flagged safety problems.
Anthropic enters financial adviser AI market
AI company Anthropic is expanding into software for financial advisers, according to Sifted.
Florida seeks to halt OpenAI's AI development over safety
Florida's attorney general has asked a court to bar OpenAI from developing new AI models without independent oversight, citing alleged…