Cute AI Agents Are Marketing Weapons That Mask Privacy Risks

By Tom Whitfield

Meta's Muse agent can send your email, shop, book travel and act inside your apps — and it greets you as Jolly, a digitized teddy bear that blinks and waves. Researchers Ankolika De of Vanderbilt University and Jessica Maddox of the University of Georgia argue that this disarming packaging is precisely the point: cuteness obscures the invasive potential of AI agents.

What are Muse, Charm and Dots?

Meta launched Muse in September 2026 as a personal AI agent reachable from a phone or computer. Alongside it, the company released Muse Charm, a pocketable device with a screen the size of a flattened golf ball that keeps Jolly accessible at all times. Journalists have called Muse "magical" and "cutesy," and compared Charm to a Tamagotchi, the handheld digital pet that swept the 1990s.

OpenAI followed with Dots, personal assistants rendered as brightly colored, globular mascot faces that float across the screen and converse with users. OpenAI says Dots will "bring you work done the way you would do it, sometimes before you even think to ask."

Startups are making the same bet. Pickle, a personal AI app embodied in a floating purple creature with large, inquisitive eyes, promotes itself as "something alive you grow attached to, that remembers your life and eventually acts alongside you."

Why does cuteness raise the stakes?

Cute design is not inherently sinister, the researchers write. It makes unfamiliar technology feel accessible and comforting. But it can also soften how users perceive extensive data collection and surveillance, and distract them into granting agents authority to act without clear consent.

The precedent differs from toys. Hello Kitty, Labubus and Tamagotchis traded on emotional attachment, but they were collectibles that could not read messages, access accounts or act on an owner's behalf. When the cute thing is an AI agent, the stakes rise sharply.

De and Maddox, who have spent years studying how companies talk about technology, say this framing can "influence perceptions of safety" and make dangers less visible. They dub the tactic "cutewashing": masking privacy concerns to drive mass-market adoption.

How does the 'cute economy' make money?

Big Tech has long leaned on a virtual menagerie — the Firefox fox, the former Twitter bird, GitHub's octopus, Mailchimp's monkey, Hootsuite's owl. Perceived innocence softens how people judge the consequences of innovation.

The researchers cite what scholars call the "cute economy." Feelings generated by cuteness build brand loyalty, sustain engagement and influence purchases. Research shows people confide more in a friendly character than a plain tool, sharing moods, routines, relationships and preferences — and granting access to messages, calendars and accounts. That yields far richer data than a routine chatbot collects.

Cuteness also lowers skepticism. Users become more likely to grant broad permissions, less likely to read the fine print, and — once attached — harder pressed to leave the agent at all.

What has already gone wrong?

The risks are not hypothetical. According to reports cited by the researchers, Muse has read a user's text messages and given out another user's home address without their knowledge.

De and Maddox say companies should clearly explain what an agent can access, what it can do and who benefits. Disclosures and privacy policies, they argue, should be written in plain language rather than dense legalese.

Their advice to consumers: before adopting an agent, ask what data it can access and what it can do with that access. As Muse, Charm, Dots and their imitators scale, the marketing question for the industry is whether trust built on a teddy bear's face survives the first breach of it.