SetupAI
EN
來源較早收集於 60m

Meta 從智慧眼鏡中移除臉部辨識程式碼

Meta 從智慧眼鏡中移除臉部辨識程式碼
PostLinkedIn
⚛️閱讀原文： Ars Technica
#privacy#wearables#biometrics#regulatorymeta-smart-glassesmeta

💡穿戴式 AI 的隱私疑慮日益升溫；了解 Meta 如何處理公眾對生物辨識功能的審視。

⚡ 30 秒速覽

有什麼變化

在 Meta 智慧眼鏡的韌體中發現了臉部辨識程式碼。

為什麼重要

這凸顯了穿戴式 AI 裝置在隱私權方面面臨的嚴格審視。開發者在消費級硬體中實作生物辨識功能時，應預期會面臨更嚴格的監管與公眾反彈。

下一步行動

檢查產品的韌體與本地儲存空間，確認是否包含任何可能引發隱私疑慮的未使用或敏感生物辨識處理模組。

誰應關注：Developers & AI Engineers

關鍵要點

  • •在 Meta 智慧眼鏡的韌體中發現了臉部辨識程式碼。
  • •Meta 在被公開發現後的 24 小時內移除了該程式碼。
  • •官方尚未針對該功能的用途或未來規劃發表任何聲明。

🧠 深度解析

背景與延伸：來自公開資料，非原文內容。引用 18 個來源。

🔑 增強重點摘要

  • •The facial recognition code, internally dubbed "Name Tag," was discovered by Wired on June 4, 2026, within the Meta AI companion app, which is essential for pairing Meta's Ray-Ban Meta and Oakley Meta smart glasses to a user's phone.
  • •The dormant code was designed to convert faces captured by the smart glasses' camera into unique biometric signatures, or "faceprints," which would then be stored on the user's device and cross-referenced with new scans to identify individuals.
  • •Meta's Vice President of Communications, Andy Stone, described the feature as a "pilot effort" and stated that no "final decision" had been made, despite the code being present in an app downloaded over 50 million times.
  • •The removal of the code on June 5, 2026, occurred less than 48 hours after public discovery and followed calls from over 70 advocacy organizations, including the ACLU, urging Meta to abandon facial recognition plans for its smart glasses due to significant privacy and civil liberties concerns.
  • •This incident follows a history of legal and financial repercussions for Meta regarding facial recognition, including a $650 million settlement in 2021 for violating the Illinois Biometric Information Privacy Act and a $1.4 billion settlement with Texas in 2024.
📊 競品分析▸ Show

A Markdown table comparing this with competitors (Feature/Pricing/Benchmarks). Return null if not applicable (e.g. op-ed, interview, single-product announcement with no clear competitors).

🛠️ 技術深入

  • The "Name Tag" system was reportedly designed to utilize three distinct AI models: one for detecting faces, another for aligning and repositioning the detected faces, and a third for converting these faces into usable biometric data or "faceprints."
  • This biometric data was intended for local storage on the user's smartphone, with the Meta AI companion app configured to receive updates from Meta's servers.
  • Faces that the system failed to recognize were designed to be cropped, indexed, and stored locally in a "pending" folder for potential future identification.
  • The Meta AI app, which serves as the primary interface and connectivity hub for Meta's smart glasses, was the platform where this facial recognition code was embedded.
  • Meta's smart glasses, including the Ray-Ban Meta and Oakley Meta models, are equipped with a 12 MP camera capable of capturing photos and videos.
  • While Meta's system aimed for on-device processing of faceprints, other smart glasses like Rokid's also feature on-device object detection models, contrasting with many devices that offload AI functionalities to connected phones due to power constraints.

🔮 前景展望基於引用來源的 AI 分析

Meta will likely proceed with extreme caution on future biometric features in smart glasses.
The swift removal of the code and the significant public and advocacy group backlash highlight the intense privacy concerns and regulatory scrutiny, compelling Meta towards a more transparent and strictly opt-in approach for any future biometric implementations.
The incident will intensify calls for stronger biometric privacy regulations for wearable technology globally.
The discovery of dormant facial recognition code in a widely distributed consumer app underscores the potential for covert surveillance and will likely fuel legislative efforts to establish clearer legal frameworks and consent requirements for biometric data collection by wearables.
Competitors in the smart glasses market will increasingly emphasize their privacy-by-design features and transparent data handling.
To differentiate themselves and avoid similar controversies, other smart glasses manufacturers will likely highlight their commitment to user privacy, explicit consent mechanisms, and the absence of covert biometric collection capabilities in their marketing and product development.

⏳ 時間線

2010
Facebook rolled out a tagging suggestion feature using facial recognition.
2021
Meta (then Facebook) shut down its platform-wide facial recognition system and deleted over 1 billion faceprints following legal backlash and a $650 million settlement.
2021
Meta debated adding facial recognition to the first generation of its Ray-Ban smart glasses but decided against it due to privacy concerns.
2024
Texas settled a $1.4 billion lawsuit against Meta for collecting and storing biometric profiles of Texan residents without consent.
2025-02
The New York Times reported on an internal Meta memo from 2025 revealing plans to add facial recognition, internally known as "Name Tag," to its smart glasses.
2026-01
Meta reportedly began quietly adding facial recognition components (NameTag) to its Meta AI companion app over multiple updates.
2026-06-04
Wired published a report uncovering the dormant "Name Tag" facial recognition code in the Meta AI app for its smart glasses.
2026-06-05
Meta released an update removing the facial recognition code from the Meta AI app.
⚛️閱讀原文： Ars Technica
📰

AI 週報

閱讀本週精選 AI 大事摘要 →

👉相關動態

同主題

探索 #privacy

同產品

更多關於 meta-smart-glasses

同來源

最新來自 Ars Technica

AI 策展新聞聚合。所有內容版權歸原始發布者所有。
原始來源： Ars Technica ↗

這是摘要，不是原文。去看原站，或訂閱每週簡報。

每週電子報

每週一封，可隨時退訂。