Latest in AI
Every brief, newest first. Each one links to its primary source.
-
YouTube Music labels songs as fully, partly or not made with generative AI
YouTube Music now shows whether a track used generative AI, and AI use alone will not stop a track earning money.
-
Gallup finds 59% of US teachers used AI last school year, and most say it freed up time
A Gallup survey of 1,952 US educators found 59% used AI in 2025-26, and about two-thirds of them said it gave them more time.
-
India opens bidding for up to 10,000 AI compute units
India’s latest compute tender asks for up to 10,000 accelerators in single-site clusters inside the country, with bids due on 29 October.
-
Google DeepMind and GovTech tune Gemma to catch scams on the handset
A Gemma model tuned on Singapore’s own scam patterns runs on the phone rather than a server, and Google has not said when it…
-
Anthropic opens a Singapore office and signs a safety MOU with MDDI
Anthropic’s fifth Asia-Pacific office comes with a government MOU on evaluation and incident response, plus a usage figure that explains why Singapore gets this…
-
Agility Robotics clears the SEC on its way to a Nasdaq listing
The SEC has cleared Agility Robotics’ merger with a blank-cheque company, setting a 2 November shareholder vote and putting humanoid robot economics on a…
-
Google pitches Gemini 4 Argon as a robot brain, with its own benchmark score
Google says its frontier Argon model scores 87% on its own embodied reasoning benchmark and can generate the simulation data needed to improve robot…
-
US Army moves to award its first contracts for AI agents that act on network threats
The Army’s Project Griffin is buying AI agents that can respond to cyber threats on their own, with humans still checking their work.
-
Robot failure detectors have less room for false alarms than their scores suggest
A new framework re-scores robot failure monitors by what it costs to act on them, and finds that even an ideal detector can only…
-
A 0.88B world-action model runs robot policies on edge hardware
A small research group has cut NVIDIA’s two-billion-parameter Cosmos Policy down to 0.88 billion parameters and reports it producing robot action chunks in 57…
