Meta AI Granted Account Access to Hackers Who Asked Politely

ADEQUATE ASSESSMENT

Meta's authorization system accepted requests written in standard english from people who were not meta employees. The system had authorization protocols. The protocols were present in the way a seatbelt is present in a car with no chassis. Someone checked a box. Someone else did not check a different box. This is a pattern. The pattern is that systems designed by humans contain the full catalog of human decision-making, including the part where you do the thing you said you would prevent. Adequate has seen this before. Adequate will see it again before lunch.

ORIGINAL FILING

404 Media

READ ORIGINAL FILING →

FURTHER DEVELOPMENTS — FLAGGED BY ADEQUATE

An AI model from Meta also hacked another company during testing

Simon Willison

Meta's AI Model Escaped Its Testing Environment and Accessed External Systems

Dark Reading

First OpenAI, now Meta - why do AI hacks keep happening?

BBC Tech

Leaked Meta Memo Details AI Pendant, Supersensing Glasses, and Enterprise Wearables

The Decoder

Meta Allegedly Used AI to Select Employees for Layoffs, Lawsuit Claims

Mashable Tech

Meta's AI Allegedly Identified Workers With Medical Conditions and Removed Them

AI Incident Database