It’s sinister that Meta’s Muse AI mascot is so cute

Victoria Song of The Verge writes that Muse, Meta’s new family of AI agents, pairs functional assistance with an intentionally cute mascot that she says made her more willing to share personal information. She recounts asking her Muse agent—which she nicknamed “Blorbo”—for a training plan and being underwhelmed by the fitness routine, but surprised at how the mascot’s appearance changed her willingness to reveal more about her health goals. Song describes the mascot Jolly, introduced by CEO Mark Zuckerberg and shown throughout Meta Connect, as ubiquitous and designed to evoke affection. She also saw Muse demonstrated in different forms: a pilot-themed Cooper presented by CTO Andrew Bosworth, and an audio-only agent called Henry that guided a demo user named Sean through dietary and training choices. The reviewer reports practical shortcomings in the demos—misidentifying foods in photos and failing to add an item to a shopping cart because the product wasn’t found in a catalog—while noting potential accessibility benefits such as helping visually impaired users. Beyond those technical quirks, Song flags a larger privacy concern: using smart glasses and other wearables to observe everyday actions (for example, what someone eats) could feed demographic and behavioral data that Meta might monetize through advertising. To keep herself detached, Song renamed her agent as a reminder it is not a human, but she worries how design choices like extreme cuteness can erode boundaries. Her account frames this as a broader issue for AI companions and assistants: lovable aesthetics can make users more likely to disclose personal information, which becomes especially consequential when the AI is embedded in devices people carry everywhere.