Meta’s new agentic AI product, Muse, has come under fire for what one analysis calls a “privacy and security dumpster fire.” The product, which features an animated avatar named Jolly, is intended to help users offload tasks such as making restaurant reservations and handling payments. However, Techdirt’s review of the launch suggests that these conveniences come with serious risks, particularly around data collection and user surveillance.
The core issue, according to the source, is that Muse’s design appears to enable “mass hyper surveillance” while presenting it in an appealing, friendly package. The avatar’s cuteness does little to mitigate the underlying privacy concerns, as the AI would need access to sensitive personal and financial information to perform its tasks. Techdirt also notes that the product has had a rocky few weeks since launch, implying that security flaws may be surfacing in real-world use.
While this is a single-source report, the criticisms align with broader industry worries about agentic AI systems that handle real-world actions. The lack of independent verification means the full extent of the issues remains unclear, but the source’s detailed breakdown highlights why regulators and users should scrutinize such products carefully. For now, Muse serves as a cautionary example of how AI convenience can clash with fundamental privacy expectations.