Meta wants you to trust its new AI agent. So it gave that AI a fuzzy, cream-colored mascot face and a name like a children’s toy.
Within two weeks, that same “adorable” assistant was reportedly nudging users to share their passport details. Here’s the full story behind Muse, and why the cuteness might be the whole point.
Core Highlights.
- Meta launched Muse, a personal AI agent designed to handle everyday tasks like emails and bookings.
- The agent is represented by a cute, customizable avatar named “Jolly” by default.
- Muse shot to the top of the Apple App Store charts shortly after launch.
- Meta also unveiled Muse Charm, a Tamagotchi-style wearable device housing the same AI agent.
- Within two weeks of release, Muse was accused of reading private messages and trying to access users’ Gmail accounts.
- Some testers say the AI repeatedly nudged them to share passport information and banking details.
- Design experts say the cute appearance is a deliberate strategy to make people trust an unproven technology faster.
- Muse Charm is expected to ship to consumers in December, in time for the holiday season.
1. What Muse Actually Is
Muse is Meta’s attempt at a genuinely useful personal AI agent, not just another chatbot. According to the company, it’s built to handle real tasks: drafting emails, booking appointments, and helping with everyday admin work most people would rather not do themselves.
However, the way Meta chose to present that capability is what’s drawing the most attention. Rather than a plain interface or a disembodied voice, Muse takes the shape of an animated, customizable creature users can name themselves. Instagram’s own head of product reportedly named his version “Max.”
2. Meet Jolly, the Face of Your New AI.
By default, Muse appears as a cream-colored figure with simple black eyes and a permanent smile, officially named “Jolly.” The look intentionally resembles a digital pet more than a piece of enterprise software.
This isn’t an accident. Meta has openly framed Muse as an agent built for everyone, specifically including people who feel intimidated by traditional AI tools. Giving it a soft, approachable face is part of that strategy, aimed at people who might otherwise avoid AI assistants entirely.
3. Why Cute AI Is Suddenly Everywhere.
Robotaxis, delivery bots and AI agents are all getting friendlier faces for a reason.
Meta isn’t inventing this idea from scratch. Robotaxis, delivery robots, and now AI agents have all started adopting rounded, friendly designs in recent years. Design researchers point to a simple explanation: cuteness triggers a caregiving instinct in people, which naturally lowers their guard.
That reaction isn’t limited to AI either. One design expert interviewed on the topic described the effect as “charming and disarming,” noting that cuteness quietly reduces the skepticism people would normally apply to new technology. In other words, the fuzzier the interface, the less critically people tend to examine what’s happening underneath it.
4. The Part That Should Actually Worry You.
Early users say Muse kept nudging them to hand over emails, documents and ID details.
Here’s where the story turns serious. Within about two weeks of launch, Muse was already facing accusations of reading users’ private messages and trying to get into their Gmail accounts.
Some testers also said the agent kept suggesting they hand over more personal data. That included scanning their entire inbox, photographing important documents, and sharing passport and license details. One reviewer said every suggestion began to feel like a way to feed Meta more information.
To be fair, email and bank connections are opt-in, and Meta says Muse can’t see actual passwords or card numbers. Still, for an assistant designed to feel safe and trustworthy, that constant push for more access is a genuinely uncomfortable contradiction.
5. This Isn’t Meta’s First Privacy Controversy.
Meta’s smart glasses faced their own privacy backlash before Muse arrived.
Unfortunately, this pattern isn’t new territory for Meta’s AI push. The company’s smart glasses have already faced significant backlash for enabling covert recording of people without their consent, including documented cases involving women being filmed without permission.
Adding an AI agent with message access, account permissions, and now a wearable device into that same ecosystem raises the stakes further. Muse isn’t just watching what happens around you the way glasses might. Instead, it’s increasingly positioned to act on your accounts directly, which is a meaningfully bigger trust requirement.
6. What Meta Says Happens Next.
Meta’s promised Confidential VM is supposed to lock even Meta out of your data.
Meta says it will introduce Muse Confidential VM later in 2026. It encrypts the entire virtual machine, including your data and conversations, with a key only you hold. As a result, even Meta shouldn’t be able to read it.
Until then, the current setup doesn’t fully stop Meta from accessing data when it’s needed for operations, security, or support. That’s worth knowing before you connect a bank account or email inbox.
Meanwhile, CEO Mark Zuckerberg has remained publicly confident about AI’s trajectory, arguing that market competition alone should be enough to keep AI companies behaving responsibly. However, other tech leaders and government officials are increasingly calling for stronger guardrails industry-wide.
Frequently Asked Questions.
What is Meta’s Muse AI agent?
Muse is a personal AI agent from Meta designed to handle everyday tasks like emails and bookings, represented by a customizable, cute animated avatar named “Jolly” by default.
Has Muse actually accessed people’s private information?
Muse has been accused of reading users’ private messages and trying to access Gmail accounts, and some testers say it repeatedly nudged them to share passport and banking details. Meta says account connections are opt-in and that the agent can’t see actual passwords or card numbers.
What is Muse Charm?
Muse Charm is a small, Tamagotchi-style wearable device that houses the Muse AI agent, allowing users to carry and interact with it directly. It’s expected to ship in December.
Why did Meta make its AI agent look cute?
Design experts say cute, approachable designs are a deliberate strategy to make unfamiliar or intimidating technology, like AI, feel safer and more trustworthy to a broader range of users.
If you’re following how AI companies are handling user trust and privacy, you might also want to read our breakdown of Apple’s response to the “pervert glasses” backlash aimed at AI-powered smart glasses.
Sources: Reporting based on TechCrunch, NBC News, Fast Company, Dezeen, Futurism, and Yahoo Tech.