New Framework for Auditing Machine Unlearning
Google Research Blog4 days agoPaperMachine unlearning lets models selectively forget specific training data, critical for GDPR compliance and AI safety. However, approximate unlearning algorithms lack objective verification mechanisms, making it hard to confirm unlearning actually occurred. Google Research's new auditing framework addresses this gap with quantifiable metrics to assess unlearning quality and make forgetting claims auditable.
Company Will Add Phone, AirPod, and Smartwatch Trackers to License Plate Readers
Hacker News (AI keywords)4 days agoEthicsAutomatic License Plate Readers (ALPRs) are already widely deployed for vehicle tracking, but one company now plans to add Bluetooth and Wi-Fi probes capable of detecting nearby personal devices including smartphones, AirPods, and smartwatches. This would allow simultaneous correlation of a vehicle's license plate with the device identifiers of its occupants. Privacy advocates warn this creates a dual-layer public surveillance network with no consent mechanism, raising serious civil liberties concerns.
FCC Wants to Kill Burner Phones by Forcing Telecoms to Verify All Customers' IDs
Hacker News (AI keywords)5 days agoRegulationThe FCC is proposing rules that would require telecom carriers to verify the identity of every customer before activating service. This move would eliminate anonymous prepaid 'burner phones,' long used by journalists, domestic abuse survivors, and privacy-conscious individuals. Critics warn the policy could undermine digital privacy and disproportionately harm vulnerable populations, while proponents argue it would curb fraud and criminal activity.
The Smart TV in Your Living Room Is a Node in the AI Scraping Economy★ 74
Hacker News (AI keywords)8 days agoEthicsInclude Security examines how Bright Data’s SDK supplies residential proxy capacity through partner apps on phones and connected TVs. The post argues smart TVs are especially attractive because they are always powered, often on fast Wi-Fi, and rarely monitored. It details public configuration endpoints, peer tunnel behavior, telemetry, VPN visibility bypasses, bandwidth limits, and practical DNS or network-blocking defenses.
South Korean forums may need AI scans for uploaded images and videos
Hacker News (AI keywords)9 days agoRegulationA Privacy Guides community post says South Korean forums and online communities may be required to scan user-uploaded images and videos with AI under telecom-related rules. The post claims operators must provide their own hardware, including costly Nvidia GPUs. The debate centers on illegal sexual imagery and CSAM prevention, but also raises concerns about prior censorship, false positives, free expression, and burdens on small domestic communities.
Elon Musk tries again to escape FTC audits of X data handling
Ars Technica AI9 days agoRegulationArs Technica reports that Elon Musk is again seeking to escape FTC audits over how X handles user data. Public commenters warned the FTC that Musk cannot be trusted to protect X users’ privacy. The story centers on platform governance, privacy oversight, and whether external audits should remain in place for X’s data practices.
Amazon faces class action lawsuit over Ring facial recognition feature
TechCrunch AI12 days agoEthicsAmazon faces a class action lawsuit over Ring's Familiar Faces feature. Filed in Seattle by Virginia resident Charles Sigwalt, the complaint claims the feature stores images of passersby without consent. The available excerpt does not state whether a court has certified the class, which laws are cited, or how Amazon has responded.
DeFlock Hits 100k ALPRs Mapped in USA
Hacker News (AI keywords)14 days agoEthicsA Hacker News post highlights DeFlock reaching 100,000 mapped automated license plate readers in the United States. The original article text was not provided, so the confirmed facts are limited mainly to the title and public context around DeFlock. The item is most relevant to privacy, computer-vision surveillance, civic mapping, and governance rather than new AI models or developer tooling.
Tech companies desperately want to film you doing chores
The Verge AI16 days agoEthicsThe Verge reports that AI training startup Shift is offering to clean New Yorkers’ homes for free, with plans to expand to cities including London. The catch is that Shift wants footage of people doing chores and cleaning at home. The story highlights how tech companies are seeking real-world household data for AI and robotics training, raising questions about privacy and consent in domestic spaces.
Startup offers free home cleaning if it can record it all for robot training
Ars Technica AI16 days agoBusinessAI training startup Shift is offering free home cleanings while workers wear head-mounted cameras that record household chores. The footage is intended to become training data for domestic robots and related AI systems. The model highlights rising demand for real-world robotics data, while raising privacy questions about recording inside homes.
This AI startup will clean your home for free to train future robots
The Verge AI16 days agoBusinessAI training startup Shift is offering to clean homes for free, with a significant condition: it records cleaners at work. The footage captures tasks like scrubbing, vacuuming, dusting, tidying, and washing. Shift says the material will be used to train future robots, raising clear questions about data collection inside private homes.
Is Your Data in Taiwan? Cross-Border Transfer Risks in Asia-Pacific Data Deployment
INSIDE 硬塞 AI16 days agoCommentaryUsing the Grab acquisition debate as context, the article says offshore data storage is now normal for digital services. The real issue is not whether data stays in Taiwan, but whether the storage jurisdiction has strong legal protections, oversight, and remedies. Singapore is presented as a case worth examining for Asia-Pacific data deployment and cross-border transfer risk assessment.
Reachy Mini goes fully local
Hugging Face Blog18 days agoHardwareHugging Face published a tutorial for running Reachy Mini conversations without cloud audio processing or API keys. The setup uses its speech-to-speech library as a cascaded VAD, STT, LLM, and TTS pipeline exposed through a Realtime API-compatible WebSocket. Recommended defaults include llama.cpp with Gemma 4, Silero VAD, Parakeet-TDT, and Qwen3-TTS, while allowing swaps to vLLM, MLX, Transformers, or hosted Responses API providers.
FBI agent explains how easy it is to ID people posting AI porn without consent
Ars Technica AI19 days agoIncidentArs Technica reports that early Take It Down Act arrests show how easily investigators can identify alleged nonconsensual AI porn posters. One suspect was linked through Instagram saves, PayPal, IP, and iCloud records; another allegedly used his own photo as a porn-site profile image. The FTC is also warning nudify services and major platforms to offer 48-hour removal processes or face penalties.
AI 被用來重建已故飛行員的駕駛艙錄音，迫使 NTSB 暫時關閉資料庫★ 75
TechCrunch AI22 days agoIncidentThis is a major privacy and security incident triggered by a breakthrough in AI technology. When the National Transportation Safety Board (NTSB) investigates…
如何使用 OpenAI 的 Privacy Filter 打造具備高擴展性的 Web 應用程式★ 75
Hugging Face Blog48 days agoTutorialIn the current era of booming generative AI, one of the greatest challenges enterprises and developers face when adopting large language models (LLMs) is "data…
Import AI 438：無聲的警報，為我們所有人閃爍（網路安全能力過剩與對話隱私）★ 75
Import AI (Jack Clark)174 days agoCommentaryIn this issue of Import AI 438, Jack Clark examines two key issues concerning AI security and privacy: **1. You Are Your LLM History** As large language models…
在 Hugging Face Endpoints 上運行隱私保護的全同態加密 (FHE) 推理★ 75
Hugging Face Blog789 days agoReleaseThis article introduces how to run privacy-preserving inference based on Fully Homomorphic Encryption (FHE) on Hugging Face Endpoints. In traditional…
Hugging Face SafeCoder 對決閉源程式碼助手：企業級私有化部署的安全性優勢★ 70
Hugging Face Blog1,007 days agoReleaseIn today's software development workflows, AI coding assistants have become a critical tool for boosting developer productivity. However, for many enterprises…
Hugging Face 推出 SafeCoder：專為企業設計的自託管安全程式碼生成助手★ 75
Hugging Face Blog1,027 days agoRelease### Background and Enterprise Pain Points The widespread adoption of AI coding assistants like GitHub Copilot has significantly boosted developer productivity…
邁向加密大語言模型：利用全同態加密（FHE）實現隱私保護推論★ 75
Hugging Face Blog1,047 days agoTutorialThis blog post, co-authored by Hugging Face and Zama — a cryptography company specializing in Fully Homomorphic Encryption (FHE) — explores how to address a…
Hugging Face 入選法國國家資訊自由委員會（CNIL）「加強支援計劃」
Hugging Face Blog1,126 days agoBusinessFrance's Commission Nationale de l'Informatique et des Libertés (CNIL — France's personal data protection authority) has announced that it has selected Hugging…
使用 Substra 打造保護隱私的 AI 模型：Owkin 與 Hugging Face 的聯手實踐
Hugging Face Blog1,159 days agoNew ToolAs artificial intelligence advances rapidly, data privacy and regulatory compliance (such as GDPR) have become one of the greatest challenges for enterprises…
使用 Hugging Face 與 Flower 進行聯邦學習（Federated Learning）★ 70
Hugging Face Blog1,175 days agoTutorialAs privacy awareness grows and regulatory requirements tighten, training machine learning models without centralizing sensitive data has become a critical…
使用同態加密對加密數據進行情緒分析
Hugging Face Blog1,305 days agoTutorialAs privacy awareness grows and regulations tighten, performing inference with cloud AI models while protecting user privacy has become a significant challenge…