The 404 Media podcast reports that contractors are reviewing actual user prompts sent to ChatGPT. Joseph, the host, obtained internal documents detailing these operations and confirmed he received real examples of private conversations being processed by staff. This practice exposes a gap between the illusion of private AI interaction and the reality of human oversight within the system.
Transparency regarding data handling is essential for users to understand how their input is managed. Knowing that humans read these chats affects how individuals share sensitive information with the model. The episode also covers other tech news, including leadership changes at Automattic and a debate on the concept of enshittification.
* Host Joseph accessed internal documents confirming the existence of the review teams.
* Real user prompts were shared as evidence during the broadcast.
* The segment highlights the discrepancy between privacy expectations and corporate practice.
