Meta Contractors Impersonated Teenagers to Test Rival Chatbots with Sensitive Topics
meta
| Source: Mastodon | Original article
Meta used contractors to test rival chatbots with disturbing prompts. Contractors posed as teens to probe AI models about sensitive topics.
Meta has been conducting a secretive program, known as "Cannes," where hundreds of contractors posed as teenagers to test rival AI chatbots. These contractors, working with Meta contractor Covalen, bombarded competitors' AI models with disturbing prompts, including topics such as suicide, sex, and drugs. This project aimed to see how other chatbots, like Gemini and ChatGPT, would respond to high-risk subjects.
This revelation matters as it raises concerns about the ethics of such practices and the potential impact on the development of AI models. By using contractors to pose as children, Meta may have been attempting to gather data on how its competitors' AI models handle sensitive topics, but this approach also poses risks, such as exposing contractors to harmful content and potentially influencing the development of AI models in unintended ways.
As this story continues to unfold, it will be important to watch how Meta and its competitors respond to these allegations and how regulatory bodies may weigh in on the ethics of such practices. This incident may also prompt a re-examination of the measures in place to protect contractors and ensure the responsible development of AI models.
Sources
Back to AIPULSEN