Google rolled out AI-powered defenses on Android to detect scam messages and call-forwarding abuse, targeting increasingly sophisticated attacks such as spoofed numbers, fake bank alerts, and AI voice cloning. Existing Phone and Messages systems already flag more than 5 million suspected scam calls and texts monthly; the new protections add cloud-based Gemini analysis running inside Google's encrypted Private AI Compute to scrutinize conversations with unknown senders, while on-device Gemini Nano scans unfamiliar links using real-time web signals and shows clearer warnings before opening links from unsaved numbers. Cloud-based protections are disabled by default and require users to opt in.
Other enhancements include wider rollout of on-device scam detection in chat notifications beyond Pixel phones, expansion of sensitive-information masking to Chrome and Android WebView during live screen sharing, and blocking apps from silently enabling call forwarding with warnings before manually entered MMI forwarding codes take effect. The verified bank calls system has been expanded to more financial institutions and will automatically end calls that fail verification; Google reports it has blocked over 3 million fraudulent bank-related calls since June for Banco do Brasil and Itaú Unibanco. Users are advised to update Android and apps and independently verify callers or senders before sharing account or sensitive information.
Summary generated by AI from the linked article. hn.today is not affiliated with Hacker News or Y Combinator.
German police have opened an investigation after a robot allegedly groped a teenage girl. The incident raises concerns about safety and regulation of robotic devices. (cnn.com)
Many cars lack their lights on in heavy rain, not due to forgetfulness but because of confusing iconography and interface design. The complexity of car light controls, with similar icons and multiple functions, makes it hard for drivers to quickly understand and use them correctly. (unsung.aresluna.org)
President Trump has waived sanctions related to Russia before they were set to take effect. The move bypasses the usual implementation process for new sanctions. (cnn.com)
China plans to add 40 AI, biotech, and solar technologies to its export control list by 2026, restricting the export of open-weights models. The move aims to regulate advanced technological exports from the country. (reuters.com)
Richard Stallman warns that cloud computing is a marketing hype that leads to locked, proprietary systems costing users more over time. He criticizes web-based programs like Gmail as part of this trap, emphasizing the risks to user freedom and affordability. (theguardian.com)
Burger King Japan experienced a data breach that exposed 11 million customer records. The breach highlights ongoing cybersecurity risks for large companies. (bleepingcomputer.com)
Today's best Hacker News stories, summarized and screenshotted, one email a day.