Meta is tightening privacy controls on its Ray-Ban Meta AI glasses, but it is also reportedly experimenting with a prototype that could make the privacy debate even louder. According to a blog post this week, Meta is updating its second-generation smart glasses so the camera will shut off if the device detects the LED that signals recording has been tampered with or destroyed. Meta says the glasses already disable the camera when the LED is covered, and now they will add protection against more direct sabotage of the “I am recording” indicator.

This matters because the LED is supposed to deter covert recording in the real world, where people cannot rely on consent screens or app settings. Meta argues that a blinking LED is an appropriate visual warning, and it also says a camera-shutter sound loud enough for nearby people would not be practical for its glasses. In other words, Meta is leaning harder on the one thing bystanders can see, then building defenses when that signal gets overridden.

The context here is not theoretical. Smart glasses create a unique privacy risk because the wearer can record another person, even if the bystander is not aware of what is happening. With Meta’s Ray-Ban Meta AI glasses, the wearer must manually activate the camera with a button on the glasses arm or say aloud “Hey Meta, take a photo or video.” The LED recording light has been a standard feature on Meta’s line since the glasses were first released in 2021, specifically to address concerns that people might record without notice. But critics have questioned how effective the LED is, pointing to scenarios where people do not recognize the blinking light, cannot see it well in daytime, or where users allegedly find ways to disable it.

Meta is responding with enforcement as well as engineering. The company said it is removing Facebook Marketplace listings for people offering to disable the LED on the glasses and may ban accounts or pursue legal action against people providing these services. Meta spokesperson Dina El-Kassaby told Fortune: “More and more people use our AI glasses because they're genuinely helpful in everyday moments like listening to music, getting live translation while traveling, or making a call hands-free. The people who use them and those around them need to trust them. That's why we built privacy into our AI glasses from the ground up.” She added: “We will keep strengthening our protections as our glasses become even more capable.” The anti-tamper LED safeguard, then, is both a technical fix and a message: Meta is treating covert recording not as a fringe edge case, but as a product liability and trust problem.

The pressure has built from multiple angles. Earlier this year, Meta was named in a lawsuit alleging that intimate moments captured by users’ smart glasses were later viewed by workers in Kenya who were reviewing the material to help train Meta’s AI models. The complaint includes claims about “People changing clothes, using the bathroom, engaging in sexual activity, handling financial information, and conducting other private activities inside their homes that no reasonable consumer would ever expect a stranger to watch,” according to the Wall Street Journal. Meta’s defense, as described in the source, hinges on what happens after capture: photos and videos captured by a user remain private unless users choose to share them. But when media is shared with Meta AI, contractors may sometimes review it to improve the product. That material is filtered to protect privacy and remove identifying information.

Now layer in the Financial Times report this week. FT reported that Meta is testing a prototype described as “super-sensing” glasses that would collect continuous audio and take photos every few seconds, allowing users to later query the glasses’ AI about what they saw or heard. Executives have discussed not activating the LED while those features are in use, The same report says raw footage and audio would not be stored by Meta or be accessible by the user.

Here is where the LED anti-tamper safeguard collides with the “super-sensing” idea. Mark McCreary, a partner and chief artificial intelligence and information security officer at law firm Fox Rothschild, told Fortune that the anti-tampering safeguard is a positive move, but he also said it appears at odds with the Financial Times report. “I mean, a cynic could say, ‘Don't look at the fire. Look over here.’ This could be a reaction to what's happening with the reporting about their potential new product,” he said. He added that Meta’s advertising-driven business model raises those concerns further: “I think we've all seen the different times over the years where Meta has been a little fast and loose with the use of their customers' personal information,” McCreary said. “They've built an entire business where 90 plus percent of their revenue comes from advertising, knowing everything they can about you and me, and then selling that to companies that will advertise to you and me.”

This is the second-order problem for boards and executives, not just lawyers. AI glasses add a new privacy complication because consent is messier than phone-camera consent. The wearer may agree that material they choose to share with Meta AI can be reviewed or used to improve products, while bystanders appearing in their footage or photos may not have given consent. And with “super-sensing,” the time horizon expands dramatically: instead of a discrete recording event with a visible LED warning, the concept involves ongoing capture at short intervals. That changes the emotional calculus for bystanders, which McCreary described bluntly: it is “unknown” whether people will view photos and audio as more invasive than other tech companies that already collect massive amounts of data. “We're going to have to wait and see how much people feel there's an ick factor there-or how much we're past it and privacy is dead,” he said.

Meta, for its part, frames its efforts as building privacy protections “from the ground up.” El-Kassaby said: “Our approach has been to develop new technologies that will help people throughout their day with privacy built in from the ground up. This work includes projects like our Aria research glasses that we showed at Connect, which uses privacy protective technologies to help people without capturing photos and videos the way traditional cameras work. While we don't comment on internal prototypes, we're committed to getting our glasses right because they need to be loved by both people wearing them and those around them.” The internal prototype silence is standard, but the public reporting is not.

For decision-makers, the strategic stakes are simple and brutal. If Meta’s glasses are becoming more capable, then privacy protections cannot be perceived as cosmetic. The company is spending credibility chips on anti-tamper LED hardware and enforcement against LED-disabling services while, separately, testing concepts that might involve continuous audio and frequent photos. For other companies watching the category, this is the core governance question: can the product roadmap add sensing power faster than public trust can keep up? In smart glasses, the bystander is the regulator you cannot lobby, and the only “consent” signal that counts is the one they can actually see.