Tech News
← Home  ·  All topics

Privacy

306 GoKawiil briefs on this topic

Anthropic and OpenAI launch rival AI finance tools with differing privacy defaults

Anthropic's Claude Personal Finance hub excludes user chats from model training by default and erases data once a bank account is unlinked. OpenAI's ChatGPT Finances instead follows users' existing model-training preferences, requiring them to manually opt out via Data Controls, and deletes synced account data within 30 days of disconnection. Anthropic also says its tool carries no ads or financial products, while OpenAI is testing ads on cheaper tiers and plans to expand ChatGPT Finances through a partnership with Intuit.

Researchers detail macOS daemons that build a personal relationship graph from device data

Security researchers examined two undocumented macOS background processes, knowledgeconstructiond and intelligenceplatformd, which have run on Macs since 2022 and build a queryable knowledge graph of a user's contacts, calendar, photos, messages, mail, location and screen-time data. Apple's only public documentation is a terse man page, and the researchers say the system can infer kinship and relationships between people, a feature not previously described publicly. They note Apple published a 2023 research paper on the underlying graph-construction techniques but has not disclosed how it operates inside macOS or exactly what data it collects and infers.

Voters reject over 40 data center ballot measures in midterm elections

Democrats regained control of both chambers of Congress in this week's midterm elections, with AI and data centers emerging as prominent campaign issues. Across the country, voters rejected more than 40 local ballot measures that would have supported new data center construction, reflecting concerns over the facilities' water and electricity use. Privacy advocacy groups, including the Electronic Privacy Information Center, framed the results as a rebuke of Big Tech's influence despite heavy campaign spending by tech figures like Elon Musk.

At least 10 local governments cut ties with Flock Safety surveillance cameras

Since this summer, more than ten U.S. municipalities and police departments—including Asheville, Richmond, Lansing, Monroe, and League City—have voted to end, restrict, or phase out contracts with Flock Safety's license plate readers and other surveillance tools. The moves follow local pushback over privacy, data-sharing practices, and procurement concerns, with Flock cameras currently deployed in over 100,000 locations nationwide.

Google to disable AI features by default in K-12 Search accounts next year

Google announced on Nov. 3 that AI tools integrated into its Search product will be turned off by default for K-12 students starting next year, requiring teachers or administrators to manually enable them. Google education VP Chris Phillips said that because search and AI are tightly linked, schools that block the AI tools will also lose access to Search entirely for students, triggering a notification explaining the restriction. In response, CNET highlighted several free alternative search engines, including Brave and DuckDuckGo, that parents and educators could use instead.

Postman accused of auto-installing AI plugin that exfiltrates API data

A report claims Postman's client silently installed an AI-related plugin without explicit user consent, and that the plugin attempted to transmit all API request and response data elsewhere. The claim surfaced via a developer changelog-style post rather than an official Postman announcement, and no response from Postman is cited.

Users report OpenAI's Codex silently re-enabled model training toggle

A developer reported that OpenAI's Codex tool had its 'Improve the model for everyone' data-sharing setting turned back on without their knowledge, despite having previously disabled it. Coworkers confirmed the same issue, and the poster noted the setting can be checked and disabled via OpenAI's data-controls and privacy pages. Other users on Hacker News had flagged similar behavior before.

Study finds Vizio TVs keep sending tracking data after opt-out

Northeastern University researchers tested 12 smart TVs across 10 brands and found Vizio's SmartCast continued sending the same volume of traffic to an Amazon Kinesis server linked to automatic content recognition even after users opted out, in all 30 test runs. Samsung's own streaming service showed similar behavior, while LG, Fire TV, VIDAA and Sony's Google TV stopped sending data once users opted out. Vizio's free WatchFree+ service also sends unencrypted show and channel names via DNS lookups to partner Amagi, with no option to disable it.

AI wearables promising to record every conversation raise privacy alarms

Several companies are now marketing AI-powered wearable devices designed to continuously capture and recall conversations, pitched as tools to overcome forgetfulness. Privacy advocates cited in the report warn that constant recording by these gadgets could alter the basic dynamics of human interaction.

Northeastern study finds Vizio, Samsung TVs keep tracking after users opt out

A Northeastern University study ran 5,943 experiments across twelve smart TV platforms and found that automatic content recognition (ACR) tracking often continues even after viewers disable it. Vizio's SmartCast devices sent data to an Amazon Kinesis endpoint in 100% of opted-out test scenarios, while Samsung Tizen devices showed ACR-like traffic in over 60% of cases; only four operating systems fully honored opt-out requests.

Ford pledges not to sell driving data as GM settles data-sharing claims for $220 million

Ford has publicly committed not to sell customers' driving data, a pledge issued as General Motors agreed to pay $220 million to settle claims it shared driver data without adequate consent. The announcements come amid broader scrutiny of automakers' data collection and sharing practices.

Google's Gemini accesses Gmail content by default; users can opt out

Google has enabled Gemini AI features by default across Gmail, including email summaries, to-do list creation, reply drafting and AI search. These features require Gemini to read email content to function, though Google maintains the data is used only for those specific tasks rather than broader model training. Users can disable the integration through account settings if they prefer Gemini not to access their inbox.