Meta's dedicated AI app is a privacy disaster, as are its facial recognition glasses

An article in TechCrunch reveals that concerns are mounting over the privacy and data protection implications of Meta's new standalone Meta AI app, which was launched recently to compete with ChatGPT. The article highlights that users are seemingly unaware that their conversations with the chatbot are being made public to a global audience. Users have the option to click a share button, leading to the widespread public exposure of ostensibly private text conversations, audio clips, and images. 

Currently, publicly shared content ranges from bizarre inquiries to highly sensitive personal information. Examples observed include users seeking advice on tax evasion, inquiring about family members' proximity to white-collar crimes, requesting character references for employees facing legal issues (including full names), and even sharing home addresses and court details. 

This situation presents a significant privacy nightmare, as Meta AI does not clearly indicate privacy settings or the public nature of posts. Furthermore, if a user's linked Instagram account is set to public, their Meta AI conversations are public by default. TechCrunch argues this flaw stems from the misguided assumption that users would desire to view other users' AI conversations. Despite Meta's substantial investment in AI, the app has seen limited downloads. However, the public sharing of such sensitive queries risks escalating into a major public relations crisis.

TechCrunch has reached out to Meta for comment but has yet to receive a response on the record.

Meanwhile, the Financial Times (£) reports that Meta's AI-powered Ray-Ban smart glasses have emerged as a surprising success, sidestepping the privacy concerns that plagued Google Glass due to their discreet, fashionable design. While capable of capturing photos, videos, and object identification, they currently lack facial recognition. However, Meta has reportedly resumed development on this feature, which could offer convenient social benefits like instant name recall during encounters.

The prospect of integrated facial recognition raises significant privacy and data protection concerns. While beneficial for the wearer, it eliminates the choice of anonymity for individuals being identified, as facial scans could be captured and cross-referenced with online images. 

The article includes a statement by Kashmir Hill, author of Your Face Belongs To Us, who argues that most people value their everyday anonymity and that society can, and should, impose limits on powerful AI, much like speed limits for cars. 

£ - The Financial Times article requires a subscription.