Meta is reportedly advancing with plans to integrate facial recognition technology into its next-generation AI-powered smart glasses, a development that has reignited significant privacy debates.
According to a report from The Information, these new glasses, internally codenamed Aperol and Bellini as part of a broader “Project Artemis” would represent a notable shift from Meta’s previous stance of shelving such features for its Ray-Ban Meta AI glasses due to ethical blowback.
The core of this new initiative is a proposed “super-sensing” capability: an opt-in, “always on” live AI feature designed to continuously process a wearer’s surroundings and utilize facial recognition to remember encountered individuals and information.
This ambition surfaces as the tech industry intensifies its pursuit of vast user data to train increasingly sophisticated AI models. The potential for creating pervasive surveillance tools and eroding personal anonymity is a central concern for privacy advocates, especially if such “always on” sensing becomes standard. Remembering everything you see might be alluring to some, but at what cost to our collective privacy and the anonymity of those around us?
The renewed push into controversial territory is reportedly emboldened by a belief within Meta that a potentially more business-friendly Federal Trade Commission, possibly under a new administration, might be more receptive to such technologies.
This perspective was also hinted at years prior when Meta’s CTO Andrew Bosworth, stated in2021 that facial recognition for AR glasses was a consideration. “It’s really a debate we need to have with the public,” Bosworth told CNBC back then. “If people don’t want this technology, we don’t have to supply it. The product is going to be fine either way. There are some nice use cases out there, if it’s something people are comfortable with.”
“Super-Sensing” AI: Capabilities and Concerns
The envisioned “super-sensing” AI in the next-generation glasses aims to extend the operational duration of the live AI features currently limited by battery life in existing Ray-Ban models. The Information suggests that the glasses would be able to identify people and provide information about them.
While this feature would require user opt-in, Meta is reportedly considering an indicator light to signal when this advanced sensing mode is active. The effectiveness of such indicators remains a point of contention, as the small LED on current glasses, meant to signal recording, has been criticized for being easily overlooked.
This development follows Meta’s April update to its AR glasses privacy policies and aligns with its limited use of facial recognition for identity verification. The broader context includes Meta’s ongoing efforts to build what it calls a “first step toward building a more personal AI.”, as stated during the launch of its standalone Meta AI application in April. The app, powered by Llama 4 large language models, now serves as the primary interface for Ray-Ban Meta smart glasses, replacing the older Meta View app.
A History of Innovation and Apprehension
Meta’s journey with AI-enhanced eyewear has been marked by both rapid feature deployment and persistent privacy alarms. The October 2024 “I-XRAY” demonstration by Harvard students starkly illustrated how the existing Ray-Ban Meta glasses could be combined with publicly available facial recognition tools like PimEyes to identify individuals and access their personal data in real-time. The case underscored the potential for misuse, a concern that Meta’s current privacy safeguards—an indicator light and user responsibility for consent—have not fully assuaged.
The company has continuously updated its smart glasses, with features like AI-powered live translation between English, Spanish, French, and Italian, and “Look and ask” visual assistance rolling out more broadly through 2024 and early 2025.
These capabilities, which allow users to get information about their surroundings without needing a wake word, are foundational to the more advanced “super-sensing” concept.
However, the methods behind training these powerful AI models have also drawn scrutiny, with ongoing lawsuits alleging that Meta’s Llama AI relied on copyrighted books for training data.
The Competitive and Regulatory Landscape
Meta’s push for more advanced AI eyewear occurs within a fiercely competitive and data-hungry tech landscape. Companies are increasingly leveraging AI to gather granular user data, partly because, as Nature put it, they are “(after “sucking the internet dry of usable information,” as Nature put it recently).” This trend is visible in Perplexity’s development of a browser for personalized AI use and OpenAI’s exploration of a social network.
Meta itself has been working to refine its AI’s behavior, stating a goal for Llama 4 involves addressing what it described as a historical left-leaning bias in leading LLMs. This technological adjustment has occurred alongside significant policy changes, such as the termination of its US third-party fact-checking program in January 2025.
As Meta forges ahead, competitors are not standing still. Some, like Solos, are emphasizing privacy with modular designs that allow camera removal. Others, such as Baidu with its utility-focused Xiaodu glasses, and Apple with its cautious “Atlas” project, are pursuing different strategies. Google is also a key player with its Android XR platform.
The next iteration of Meta’s Ray-Ban glasses, potentially arriving in 2025 with integrated displays, will likely be a critical step in this evolving market, bridging current wearables with the more ambitious augmented reality systems like Meta’s Orion prototype.
