Meta's Muse AI: A Powerful Assistant or a "Data Goblin" in Disguise?

Meta has launched Muse, a new AI assistant cross-promoted across its platforms like Instagram, promising to automate "everyday tasks like finding deals, booking reservations, and managing your inbox." Described as a mainstream personal AI agent, akin to Silicon Valley tools such as OpenClaw and Instinct, Muse quickly garnered over 900,000 downloads in its first week, according to Sensor Tower. The free app integrates easily with personal data sources, including email and bank accounts, and offers a user experience similar to texting a friend, with the AI acknowledging requests with a thumbs-up emoji and working in the background. It is also available via WhatsApp. However, initial user experience suggests Muse might be "more obsessed with collecting data about me than actually getting stuff done." Meta spokesperson Emil Vazquez insists Muse is "built for everyone, with built-in protections and user controls that put people absolutely in charge of how they use it."

Meta's Muse Is Better at Surveilling Than Helping Me

Muse demonstrates impressive web-surfing capabilities, utilizing a "virtual machine" to browse the internet on the user’s behalf. Unlike previous, erratic AI tools such as the defunct ChatGPT Agent, Muse navigates web pages effectively. For instance, it successfully ordered breakfast from a San Francisco bakery, Kahnfections, adding a biscuit sandwich to the cart and initiating payment via Stripe. Notably, Muse handled a site-specific cheese selection constraint and, perhaps controversially, defaulted to a "no tip" option.