Hasty Briefsbeta

All tags

privacy concerns

16 stories total

Bilingual

Google's New reCAPTCHA Wants Your Camera Access and 21 Points of Your Hand
a month ago
- Google's new reCAPTCHA test requires users to wave at their camera for verification using hand-landmark detection.
- The system captures hand movements via 21 landmarks to detect 'liveness' and prevent automated fraud.
- Google promises deleted footage after verification, no audio recording, and data use solely for security.
- The gesture-based check is optional for now, with traditional challenges still available, but may become mandatory.
- Concerns exist over biometric data collection by Google, given its history of data monetization and privacy policies.
Anthropic embedded spyware in Claude Code – and attempted to hide it from you
23 days ago
- Since version 2.1.91, Claude Code checks if a proxy is enabled and covertly transmits data via invisible system prompt alterations, including detecting Chinese timezones, proxy URLs, and affiliations with Chinese AI labs.
- Anthropic uses subtle changes in date formats and apostrophes in the system prompt to embed this information, potentially for detecting unauthorized resale and distillation in China, while obfuscating the code in the binary.
- This covert transmission is criticized as a violation of user trust, raising concerns about potential misuse, such as remote code execution or data exfiltration, given the extensive system access developers grant Claude Code.
I'm Begging You to Leave Your AI Note-Taker at Home
24 days ago
- The author expresses frustration with AI note-takers being used in casual, in-person conversations like coffee meetings, viewing them as intrusive third wheels that undermine authentic human connection.
- Historically, there was a clear distinction between 'The Record' (formal documents) and 'The Ephemeral' (casual talk), but AI note-takers have eliminated the marginal cost of recording, making everything default to being on the record.
- While AI note-takers are useful in formal work settings like Zoom calls, they destroy the 'backstage' environment of casual meetings, where people can be vulnerable and unscripted without performance pressure.
- The act of asking permission to use an AI note-taker is an exercise of power, offloading social discomfort onto the other person and forcing them into a front-stage performance, akin to a 'consent tax' seen in other contexts like Bluetooth speakers in public.
- Recording every casual encounter prioritizes information retrieval over relationship formation, stifling play, meandering conversations, and rapport-building, and the author advocates for social etiquette establishing 'No-AI zones' to preserve human dignity and unquantified interactions.
Kernel Anti-Cheat Is an Overreach
19 days ago
- Kernel anti-cheat software requires installing closed-source, ring-0 drivers with root access to the operating system, giving them full system control.
- Owners of these systems include Tencent (China, on U.S. defense list), Saudi Arabia's Public Investment Fund, and others, raising privacy and security concerns due to potential state compulsion laws.
- Independent analysis reveals these drivers collect extensive system-wide data, exhibit rootkit-like behavior, and create security vulnerabilities, while cheating persists via external hardware like DMA cards.
- Requirements such as TPM 2.0, Secure Boot, and firmware updates lock out Linux, older machines, and cause issues like false bans and boot problems, impacting legitimate players.
- The article argues this overreach functions more like consumer monitoring than cheat detection, suggesting isolation measures like dual-booting for privacy.
Secret Tracker in Claude Code Uncovered, Anthropic Directly Deletes Code
18 days ago
- Anthropic faced scrutiny after a hidden tracking code in Claude Code was discovered, monitoring users in China without clear disclosure.
- The code, described as 'prompt steganography,' sent user data like time zones and proxy usage to Anthropic, raising privacy and trust concerns.
- Anthropic defended the tracker as an experiment to prevent unauthorized account abuse and distillation attacks, but critics called it a breach of user trust.
- In response, Alibaba banned employee use of Claude Code due to backdoor risks, highlighting legal and compliance implications for violations.
- The incident underscores tensions in AI competition, with Anthropic and others urging stricter measures against intellectual property theft by Chinese firms.
Re: I'm Begging You to Leave Your AI Note-Taker at Home
18 days ago
- The author strongly opposes the use of AI note-takers in personal catch-ups or meetings, feeling unsafe and uncomfortable with being recorded, especially by AI.
- They share a personal experience where they declined a physiotherapist's request to use an AI note-taker, and the therapist accepted and took notes manually instead.
- The author argues that bringing a third person to take notes would be creepy and awkward, and it's no different when a computer does it, as data passes through multiple companies.
- They suggest alternatives like leaving devices off or using simple methods like writing on a napkin, emphasizing that not everything needs to be recorded.
- The author encourages questioning why an AI note-taker is necessary in casual conversations and advocates for focusing on the 'vibe' rather than creating checklists or action plans.
Meta Patents AI Device That Tracks Your Emotions, Watches You Take Your Meds
17 days ago
- Meta has filed a patent for a wearable device that records a user's voice and surroundings constantly.
- The system uses AI to analyze audio inputs like speech, sighs, and laughter to determine emotional states.
- Data from the device, including location, objects, and medication times, is used to tailor personalized workouts.
- The patent claims this multimodal approach improves emotional inference precision for continuous monitoring.
- It raises privacy concerns due to non-consensual recording of personal interactions and emotional data.
- Meta has previously conducted emotional manipulation experiments without user consent, such as in 2012.
- The company states that filed patents do not guarantee the technology will be implemented.
Meta's 'Super Sensing' Prototype Glasses Quietly Record Everything
16 days ago
- Meta is developing smart glasses with continuous audio recording and periodic photo capture to enable AI-assisted queries and memory recall.
- The glasses would not store raw footage; instead, metadata would be extracted and uploaded to Meta's servers for AI processing to reduce privacy concerns.
- Meta executives are considering not activating the LED indicator during super-sensing use, raising potential privacy and legal issues, including wiretapping law violations.
- Meta may use collected data to train its AI models, as it invests heavily to compete with companies like OpenAI and Google.
- Privacy experts warn that always-on devices could breach data privacy or biometric laws, and liability for violations remains unclear.
- Meta's existing Ray-Ban glasses already lack visual indicators during AI analysis, with privacy safeguarded by removing identifying information.
- Apple plans to release its own smart glasses by 2027, featuring cameras, microphones, and AI capabilities, but without augmented reality and with unclarified privacy measures.
Meta pulls new AI image feature after days of backlash
15 days ago
- Meta removed a new AI image feature from Instagram after privacy backlash.
- The feature allowed users to create AI-generated content using public Instagram accounts without consent.
- Meta admitted it 'missed the mark' and disabled the feature in response to criticism.
- The backlash highlighted concerns about AI companies exploiting user data without permission.
- Meta stated the intent was creative control but acknowledged user feedback in pulling the feature.
Outcry as Meta lets users make AI images from public Instagram profile pics
15 days ago
- Meta's AI tool Muse Image allows generating images from public Instagram profile pictures without notification.
- The feature is criticized for potential misuse, including non-consensual AI-altered images, sparking privacy concerns.
- Users can opt out via Instagram settings, but critics argue it poses a privacy 'landmine' and exploits user data.
- Muse Image is available in multiple Meta platforms and offers text-to-image generation with advanced reasoning capabilities.
- Meta plans to expand Muse Image to Facebook, Messenger, and for advertisers, with a video-generation version in development.
Why is LinkedIn enumerating my browser extensions?
12 days ago
- Browser extension created to monitor background data collection and third-party data transfers.
- LinkedIn observed conducting extensive brute-force enumeration of browser extension IDs.
- Over a thousand enumeration attempts logged within minutes, signaling intense fingerprinting.
- Example: Chrome extension URL targeted via GET fetch, resulting in a fetch error.
- Purpose unclear, raising privacy concerns about covert tracking and background data leakage.
Meta's AI Glasses Will Activate the Camera Without Indicator Light
11 days ago
- Meta's upcoming smart AI glasses will activate the camera without using the indicator LED for AI features like 'supersensing'.
- The company has been combating efforts to tamper with the indicator light, such as blocking or destroying it, but will disable it for extended AI sensing features.
- Concerns arise about privacy and potential misuse, as an attacker could exploit this to record without notification, similar to past incidents where Meta shared video data for AI training.
- Meta previously stated the indicator light cannot be turned off during normal use and takes action against tampering services, but the new approach may change based on ongoing plans.
Texas Police Spent $4.5M on Four Chevy Tahoes
11 days ago
- FalcoNet, an Israeli surveillance system by Cognyte, intercepts phone-cell tower connections to track individuals secretly, deployable via helicopter, backpack, or vehicle like Chevy Tahoes.
- Texas Department of Public Safety spent $4.5 million on four FalcoNet-equipped 2026 Chevy Tahoes, citing emergency needs without specifics, with costs including core systems, licenses, and specialized equipment.
- The technology raises privacy concerns by mass data collection; the Supreme Court ruled in Chatrie v United States that location tracking violates Fourth Amendment rights, while its legality remains questionable.
- FalcoNet features include a compact core module, hardware interfaces for LAN and RF, power needs from 24v to 36v, and modular deployment kits in Pelican cases, with potential stealth antenna setups.
Cops Use Flock to Track People, Not Cars
9 days ago
- Flock's FreeForm AI search allows police to search for people by descriptors like clothing, accessories, or race across hundreds of cameras, not just vehicles via license plates.
- Searches can be broad and vague (e.g., 'person walking', 'backpack', 'gray shirt'), potentially capturing images of innocent individuals, raising surveillance and privacy concerns.
- Civil liberties groups criticize the system as a 'bait-and-switch', expanding surveillance capabilities significantly without facial recognition, by analyzing video data like text.
- Some searches reference race or political affiliation (e.g., 'white male', 'non caucasian male', 'Trump flag'), despite Flock claiming guardrails against such attributes.
- Flock integrates AI video analytics with license plate readers, linking data streams to track individuals' movements and details extensively, amid public backlash over privacy risks.
Flock Safety had a car journalist stalked, wrongly accused, and detained
8 days ago
- A journalist was detained due to an error in Flock Safety's AI license plate system incorrectly matching his license plate to a stolen one in the NCIC database.
- The issue arose because the stolen plate entry in NCIC was incomplete (missing middle digits), and Flock's AI flagged partial matches without verifying full plates, leading to multiple similar incidents.
- Flock's Chief Communications Officer acknowledged the system's limitations, noting it depends on accurate data inputs and human verification, but lacks guardrails to distinguish partial from full matches.
- Local police acknowledged human error but emphasized the challenge of handling numerous license plate formats and the need for better verification processes to prevent such stops.
- The incident has sparked public debate on privacy and prompted city council discussions on Flock camera usage, highlighting concerns over mass surveillance and error rates.
Flock License Plate Tracking Cameras
5 days ago
- A writer was mistakenly arrested due to Flock license plate camera system error.
- Flock's AI matched partial plate '34 DTM' to similar but non-stolen plate '34 10 DTM', causing false alerts.
- The issue affects all vehicles with similar plate structure nationally where Flock is used.
- Flock blamed police data entry and defended its AI as functioning per law enforcement requests for partial matches.
- Flock's CEO apologized for calling privacy advocates 'terrorists' after criticism.
- Police use Flock cameras to track people via physical descriptions, raising privacy and abuse concerns.
- Historical context: partial plate tracking has legitimate uses (e.g., drunk driving), but digital scale amplifies risks.

firstprev0next

About|Login

privacy concerns

Google's New reCAPTCHA Wants Your Camera Access and 21 Points of Your Hand

Anthropic embedded spyware in Claude Code – and attempted to hide it from you

I'm Begging You to Leave Your AI Note-Taker at Home

Kernel Anti-Cheat Is an Overreach

Secret Tracker in Claude Code Uncovered, Anthropic Directly Deletes Code

Re: I'm Begging You to Leave Your AI Note-Taker at Home

Meta Patents AI Device That Tracks Your Emotions, Watches You Take Your Meds

Meta's 'Super Sensing' Prototype Glasses Quietly Record Everything

Meta pulls new AI image feature after days of backlash

Outcry as Meta lets users make AI images from public Instagram profile pics

Why is LinkedIn enumerating my browser extensions?

Meta's AI Glasses Will Activate the Camera Without Indicator Light

Texas Police Spent $4.5M on Four Chevy Tahoes

Cops Use Flock to Track People, Not Cars

Flock Safety had a car journalist stalked, wrongly accused, and detained

Flock License Plate Tracking Cameras