In what is arguably a rare move for a company that seems to always lurch from one privacy fiasco to another, it’s just possible that Meta may have listened to some vocal critics. The company may, reports say, be planning to release a version of its controversial AI-powered smart glasses that will lack cameras.
On paper, Meta’s smart glasses endeavour looks impressive. The company’s clever partnership with Ray-Ban helped propel AI-equipped digital eyewear into the mainstream, and has reportedly led to millions of sales of these gadgets that sport microphones, speakers, cameras and – for new models – tiny built-in displays.
But then it was rumoured that Meta had plans to use the devices’ cameras for automatic face recognition on a mass scale, and had even downloaded some of the necessary code to the glasses.
The public backlash was immediate, and spread beyond the issue of face recognition to include worries about bad actors using the specs to record illicit videos without consent. To some people, these are now “pervert glasses,” a device with questionable ethics, from a questionable company, only used by people who have ill intentions. Some have even called for a ban on the devices altogether.
News site The Information has heard from Meta insiders on the matter and from what they revealed, it seems like some of this privacy criticism has hit home with Meta’s leadership.
According to the report, the company is planning to release a version of its smart glasses that will come with most of the benefits of the earlier versions, but will ship without cameras. Instead, they’ll have an array of six microphones and will have speakers like the existing models. The microphones are for accurate noise suppression and to hear the wearer’s commands, and the speakers direct sounds into the wearer’s ears, but are almost imperceptible for other people nearby. The glasses, which may be called Luna, also have a dedicated button to activate Meta’s AI chatbot. They could debut at the company’s developer conference, set for Sept 23 and 24, and they may ship in October, The Information said.
While you could read this as an admirable pivot by Meta to meet its critics halfway – a lesson for any CEO of a company with a controversial product – there are a few details to ponder.
First, when you’re designing a high-tech product like this, even if it’s a variation on an existing one, the lead times can easily stretch into years for engineering, testing, and supply chain preparations. Then consider the new device’s rumoured AI button: it’s a dead giveaway. By making it easy to access, Meta wants you to be using its AI chatbot as a default, instead of other rival systems like Apple’s Siri or OpenAI’s ChatGPT. Lastly, by omitting cameras, Meta may be able to sell the rumoured devices for less, and it’s even possible they’ll have longer battery life. These are changes which could make the glasses more appealing to potential customers.
In the long run, cam-free glasses could simply be a play for user subscription revenues and swathes of user data for training future AI models.
Meanwhile, privacy champions will note that while camera-free Meta glasses deal with one angle of the privacy problem, the microphones might remain an issue. They could be used to record conversations or other sounds, potentially in compromising or legally messy situations.
Meta, after all, doesn’t have the best reputation for data protection. In March, it was hit by controversy when the public learned that third-party companies and Meta employees had been seeing videos and photos recorded by earlier glasses versions, including in private or even intimate situations. And let’s never forget the Cambridge Analytica scandal from 2018, when users’ Facebook data was accessed in ways linked to political ad targeting.
Apple may be poised to test out similar audio-only AI features via its new Apple Watches, which have an “always listening” feature. In a coming update, Apple’s AI systems will allow new Watch users to rewind the previous 15 seconds of conversation, viewing it as a text-based transcript. There’s also a system that listens to a user’s entire day, and then provides a text-based summary, which could be handy for people who like to journal.
Apple has dodged some privacy concerns by processing audio information in a secure and anonymous way, and by making it impossible for anyone to play back actual recordings. Still, the new AI audio features are ethically challenging. Legal experts have already chimed in, suggesting that there are few precedents for this kind of tech and that some eavesdropping implications may have to be tested in court.
It seems likely that similar audio privacy concerns will hit Meta’s rumoured glasses, too.
However, as the drama around Meta’s glasses plays out, there are useful business lessons here. The public is very protective of their privacy: move too fast, and intrude too far with new features that cross people’s red lines, and your company could also face a customer backlash. Pivoting to shape your policy around people’s privacy sensitivities could repair some of this damage. But truly savvy leaders will have factored privacy issues in from the get-go, and have mitigations and protections already in place. – Inc./Tribune News Service
