The Mysterious Origins of the Most Energetic Neutrino Ever Detected

Two years ago, Earth was hit by the most energetic neutrino in history. Scientists thought it might be a measurement error, but a new study has confirmed this “ghost particle” was real. [...]

SynthID Detector can check media to see if it was generated with Google's AI tools

Google originally launched SynthID, its digital watermark for AI-generated content, as a way to detect whether an image was created using the company's Imagen model in 2023. Now, at Google I/O 20 [...]

This Giant Subterranean Neutrino Detector Is Taking On the Mysteries of Physics

Located in China, Juno is a 17-country collaboration that will try to detect neutrinos and antineutrinos to learn more about their mass. [...]

Could You Use a Rowboat to Walk on the Seafloor Like Jack Sparrow?

In Pirates of the Caribbean, Jack and Will use an overturned dinghy to hold air underwater. Madness or brilliance? [...]

NASA's Curiosity rover gets its drill stuck, recordings from the Arctic seafloor and more science stories

This week's science news. [...]

Meta’s AI detector can’t catch its own cropped fakes

The Meta AI detector promises to catch Meta’s own fakes. Crop the image, and more than half slip straight past it. The tool was meant to be a fix for the deepfake problem, not an example of it. This [...]

Pangram 3.0 AI text detector claims up to 99.98% accuracy, even for subtly AI-assisted content

Pangram has released version 3.0 of its AI text detector. The most important new feature: the tool now classifies text into four categories instead of making a simple binary distinction between human [...]

Deezer’s new free tool scans your Spotify and Apple Music playlists for AI slop

Deezer has built a free tool that will tell you how much of your music was made by a machine, even if you stream it somewhere else. The French service has opened its AI-music detector to the public, l [...]

Pangram says its new AI text detector makes only one mistake per 24,000 documents

Pangram 4 detects 99.66 percent of AI-generated text with just one false positive per 24,000 documents, the company claims. The model also resists "humanizer" tools that disguise AI writing [...]