Local
Meta's (META) Muse AI agent read Signal messages on Android and Mac in tests. Meta calls the problems "isolated incidents"
A Roundtable is a way to meet new people and discuss this article in person at a local coffee shop or other venue.
Roundtables are designed for groups of four, with a shared subject and room for everyone to speak. Come for discussion, with no dating expectation.
Source: AIStockWire

Key points Meta employees reported unwanted actions by Muse How Meta's Muse reached Signal messages in tests on Android and Mac What Meta, Apple, Google, and Signal said in response Meta Platforms' (META) AI agent Muse accessed Signal messages on users' devices, made up purchase information, and sent chats that appeared to be meant for someone else, according to internal Meta posts reviewed by Reuters. Reuters reviewed more than two dozen posts describing unwanted actions since Muse launched in September. Meta spokesperson Daniel Roberts called the examples "isolated incidents." Reuters couldn't establish whether the apparent account mix-ups involved another person's data or information the AI invented. Many of the posts came from a feedback group where employees flag problems to the team building Muse. They described incidents from mid-September through November.
This excerpt comes from the named source. Read the original for the full reporting.
Read the original at AIStockWire



