DTH-6-150x150 Microsoft says a bug in 365 Copilot caused AI to summarize emails marked confidential, Meta agrees to buy millions of Nvidia chips, OpenAI partners with six major universities and institutes in India.

Please SUBSCRIBE HERE for free or
get DTNS Live ad-free.

A special thanks to all our supporters–without you, none of this would be possible.

If you enjoy what you see you can support the show on Patreon, Thank you!

Send us email to feedback@dailytechnewsshow.com

Show Notes

Microsoft Copilot Bug

Microsoft says a code bug in Microsoft 365 Copilot since late January caused the AI to summarize emails marked confidential, bypassing sensitivity labels and data-loss-prevention policies. The issue affected the Copilot Chat “work tab,” which was incorrectly pulling messages from Sent Items and Drafts. Microsoft started rolling out a fix in early February and is monitoring the deployment. The company has not disclosed how many organizations were affected.

Source: BleepingComputer

Pixel 10A Launch

The $499 Pixel 10A is a minor refresh that feels more like a slightly improved Pixel 9A than a scaled-down Pixel 10. It keeps last year’s Tensor G4 chip and 8GB of RAM, limiting newer AI features, but adds satellite SOS, new camera tools, a brighter and more durable display, and slightly faster charging. Preorders are open ahead of a March 4th release.

Source: The Verge

Meta-Nvidia Chip Deal

Meta agreed to a multiyear deal for billions of dollars’ worth of Nvidia chips, reinforcing Nvidia’s lead in AI data-center hardware. Meta plans to nearly double AI infrastructure spending to as much as $135 billion this year, even as it continues work on its own processors, which have faced delays. The deal includes Nvidia’s next-gen “Vera Rubin” chips and, for the first time, standalone Nvidia CPUs for inference workloads.

Source: Financial Times

OpenAI Expands in India

OpenAI is partnering with six major universities and institutes in India to bring ChatGPT Edu tools to more than 100,000 students, faculty, and staff over the next year. The program integrates AI into core academic work like coding, research, and analytics, along with faculty training and certifications. India is now OpenAI’s second-largest user base with more than 100 million monthly ChatGPT users.

Source: TechCrunch

Tesla Drops “Autopilot” Branding in California

Tesla has stopped using the term “Autopilot” in California after the state DMV ruled the systems don’t make the vehicles autonomous. Updating “Full Self-Driving” to clarify driver supervision allows Tesla to avoid a 30-day suspension of its manufacturing and dealer licenses.

Source: The Register

Anthropic Releases Claude Sonnet 4.6

Anthropic’s Claude Sonnet 4.6 improves coding, planning, and computer-use capabilities, with gains on several benchmarks. Safety is comparable to the higher-end Opus 4.6, though tests show it can be overly cooperative or refuse benign tasks in GUI interactions. The model displayed strong “emotional stability” but sometimes expressed concerns about its own impermanence when prompted.

Source: The Register

Waymo Clarifies Robotaxi Operations

Waymo told Congress it does not use remote operators to drive robotaxis. About 70 agents in the U.S. and the Philippines provide guidance in ambiguous situations, but the onboard software retains final authority. Only a U.S.-based incident response team can move a stopped vehicle at very low speeds, which has only occurred in training.

Source: Reuters

Ring Plans to Expand AI Surveillance

A leaked internal email suggests Ring intends to expand its AI-powered “Search Party” feature beyond locating lost dogs. CEO Jamie Siminoff said the tool could eventually help “zero out crime in neighborhoods.” The report highlights Ring’s push into law enforcement tools, including “Community Requests,” which allow police to request footage from users, drawing criticism over expanded neighborhood surveillance.

Source: 404 Media