Microsoft hires human reviewers for Copilot AI, sparking privacy questions
Turns out, Microsoft hires real people to review what users type into Copilot AI, including stuff that gets pretty explicit.
Most users have no idea their prompts could end up being read by humans, which is raising some serious privacy questions.
The reviewers' main job is just using their gut to judge if the AI is doing what it's supposed to.
Contractors see sexualized and disturbing content
These contractors aren't just checking harmless questions. They often see disturbing material, like sexualized images or prompts involving cartoon characters in weird scenarios.
Many weren't warned before viewing the requests, leaving them with ethical dilemmas and emotional stress.
It's a reminder that behind all the smart AI features we use every day, there are real people doing tricky and sometimes uncomfortable work.