Meta pulls AI image tagging feature after backlash over privacy concerns
Meta has removed its AI image tagging feature following backlash over privacy concerns. The company's swift removal highlights growing regulatory scrutiny of AI features. The incident underscores the need for tech companies to prioritize user consent.
TL;DR
Meta unveiled Muse, a privacy‑first AI assistant, on September 8, 2026, promising isolated cloud processing and a default opt‑out from model training. Rapid adoption and a stock jump signal market confidence, while new subscription and e‑commerce models could reshape Meta’s revenue beyond advertising.
Key facts
- Muse runs in isolated cloud virtual machines with a Sentinel AI layer.
- Muse recorded roughly 2.8 million installs in its first twelve days.
- Meta’s stock rose more than 20 percent after the announcement.
- Subscription tiers are free, $20 per month, and $100 per month.
- Partnerships with Walmart and PayPal were announced at Meta Connect 2026.
Meta launched its personal AI agent, Muse, on September 8, 2026, promising strong privacy safeguards such as isolated cloud virtual machines and a default opt‑out from model training. The rollout, aimed at U.S. users aged 18 and older, quickly attracted roughly 2.8 million installs in its first twelve days and sent Meta’s stock up more than 20 percent. The company also introduced a three‑tier subscription model ranging from free to $100 per month, and announced partnerships with Walmart and PayPal to enable transactional capabilities within the agent.
Background
Meta’s entry into the AI‑assistant market comes after years of criticism over how the social‑media giant handles user data for advertising. By positioning Muse as a privacy‑first service, Meta seeks to shift its revenue model away from pure ad sales toward direct consumer subscriptions and transaction fees. The agent runs in isolated cloud virtual machines, each overseen by a Sentinel AI layer that flags sensitive tasks before execution. Users can opt out of model training by default, and Meta has pledged that Muse data will not feed its advertising systems. The company plans to test a more locked‑down version, Muse Confidential VM, later in 2026, where data encryption keys are held solely by the user. Partnerships announced at Meta Connect 2026 with Walmart and PayPal aim to embed purchasing and payment functions directly into the agent, potentially opening new e‑commerce revenue streams. The subscription tiers—free, $20 per month, and $100 per month—offer varying levels of capability, and the service is accessible via iOS, Android, web, and WhatsApp. This strategic pivot reflects Meta’s broader push into hardware like AI glasses and a desire to build an ecosystem where the AI agent acts as a central interface for a range of digital activities.
What happened
On September 8, 2026, Meta introduced Muse, a personal AI agent that operates in isolated cloud virtual machines, according to CryptoBriefing. The launch featured a default opt‑out for model training and a promise that Muse data would not be used in Meta’s advertising systems. The company offered three subscription tiers—free, $20 per month, and $100 per month—each providing different levels of functionality. Within the first twelve days, Muse recorded roughly 2.8 million installs, and Meta’s stock rose more than 20 percent following the announcement, as reported by the same source. At Meta Connect 2026, held September 23‑24, Meta announced partnerships with Walmart and PayPal to expand Muse’s transactional capabilities. The agent can send emails, shop online, and book services, and it integrates with iOS, Android, web interfaces, and WhatsApp. Meta also disclosed plans to test a Muse Confidential VM later in 2026, which would encrypt all data using keys held exclusively by the user. The Sentinel AI layer monitors permissions and flags sensitive tasks before they execute, providing an additional privacy safeguard. These moves collectively signal Meta’s shift from an advertising‑only model toward direct consumer revenue and e‑commerce facilitation.
Market impact
Meta’s stock surged more than 20 percent after the Muse announcement, indicating strong investor confidence in the privacy‑focused AI agent. The rapid adoption of roughly 2.8 million installs in the first twelve days demonstrates significant consumer interest, especially given the service’s availability across major platforms like iOS, Android, web, and WhatsApp. The three‑tier subscription model—free, $20 per month, and $100 per month—creates new recurring revenue streams that could offset Meta’s traditional reliance on advertising. Partnerships with Walmart and PayPal suggest potential future earnings from transaction facilitation, though specific financial projections were not disclosed. The planned Muse Confidential VM, with user‑held encryption keys, may attract privacy‑sensitive users willing to pay for heightened security, further diversifying Meta’s revenue base.
Why it matters
The launch of Muse marks a strategic shift for Meta, moving from an advertising‑centric business to one that includes direct consumer subscriptions and transaction facilitation. By emphasizing privacy through isolated virtual machines and a default opt‑out from model training, Meta attempts to address longstanding concerns about data exploitation. The rapid user adoption and stock price reaction suggest the market views the privacy narrative as credible and potentially lucrative. Partnerships with major retailers like Walmart and payment providers such as PayPal could embed Meta deeper into everyday commerce, positioning the AI agent as a gateway for online purchases. If Muse Confidential VM proves successful, it could set a new standard for user‑controlled encryption in consumer AI services, influencing broader industry practices around data sovereignty.
DeFi Intel’s take
Editorial analysis, not a statement of fact.
DeFi Intel's take: Meta’s Muse rollout suggests the company is betting that privacy‑first positioning can unlock new revenue streams beyond advertising. The strong stock reaction and rapid install numbers indicate investor and user appetite for an AI agent that promises data isolation. We read this as a signal that large tech firms may increasingly adopt subscription and transaction‑based models, especially when paired with high‑profile retail partnerships. The upcoming Muse Confidential VM could further differentiate Meta in a crowded AI market, potentially attracting users who prioritize encryption and user‑held keys. However, the long‑term success will depend on conversion rates across the $20 and $100 tiers and the ability to monetize e‑commerce interactions without compromising the privacy promises that initially drove interest.
Sources used for this analysis
DeFi Intel wrote this analysis from the reporting above. Any figure or quotation can be checked against those pages.
Frequently asked questions
What is Meta's Muse AI agent?
Muse is Meta’s personal AI agent launched on September 8, 2026. It runs in isolated cloud virtual machines, offers a default opt‑out from model training, and promises that its data will not be used for advertising. The agent can send emails, shop online, book services, and integrates with iOS, Android, web, and WhatsApp.
How many people installed Muse shortly after launch?
Roughly 2.8 million installs were recorded in the first twelve days after Muse’s launch, according to CryptoBriefing.
What subscription options does Muse offer?
Muse is offered in three tiers: a free tier, a $20‑per‑month tier, and a $100‑per‑month tier. Each tier provides different levels of capability, with higher tiers granting more advanced functions.
DeFi Intel is an entity-graph aggregator: we curate, tag and link crypto news to a typed knowledge graph of protocols, tokens, people and incidents. We do not republish the full article body. Use the link above to read the original report at Crypto Briefing.
Entities in this story
Discussion
Comments are reviewed before publication.
Want the full article?
Continue reading on Crypto Briefing →
No comments yet.