AM-IA brings together news, research papers and applications relating to artificial intelligence published around the world into a single feed — and makes them available in your European language at the click of a button.
Scroll to browse · Click to filter
In the 2026 edition of the Cost of a Data Break, compromised APIs and cloud configuration errors each account for 27% of the breaches that have reached an AI system.
The code assistant editor opens a repository hosting service integrated with his office client, in direct competition with GitHub.
Valued at $2 billion, the start-up adds an in-house voice recognition model, a meeting note taker and a laboratory dedicated to man-machine interfaces.
Researchers have found, in a candidate version of macOS Tahoe 26.7, a demonstration of headphones equipped with infrared cameras that power Visual Intelligence.
A column in the Journal du Net retraces the shift of Mistral: from the training of original models to the promotion, on its own platform, of the Chinese GLM 5.2 model.
Driven by a surge in sales, Anthropic achieved a positive operating result in the second quarter of 2026, the first for an AI laboratory of this size.
The young push specialized in the generation of videos and images by IA quadruples its valuation in eight months, carried by a switch to a professional clientele.
Pixxel and Sarvam prepare Pathfinder, a 200 kg satellite combining hyperspectral imaging and data center class calculation, to process images directly in orbit by the end of 2026.
Given to the Linux Foundation, the standard of communication between IA agents born at Google now demands integration with Google, Microsoft and AWS, and deployments in production.
96% of college students tried the AI tutor at least once, but the median student only wrote to him one day in three. The gain looks like that of Khan Academy without AI.
A draft letter from the State Department warns that participating in the US framework is incompatible with joining competing initiatives. Beijing responds by invoking the digital sovereignty of everyone.
The Reconstruction benchmark submits seven models to 643 papers from six disciplines, giving them only the list of references prior to publication.
The researcher dismantles the promise of the head of Anthropic by relying on a series of specialists: neither the body of knowledge nor the regulatory deadlines allow for this timetable.
▶
Revue hebdomadaire des faits marquants : la suspension du modèle Astra par OpenAI, une cyberattaque conduite par IA et la course aux 500 milliards de dollars d'infrastructure.
▶
Un point quotidien d'une minute sur l'actualité de l'IA, condensé des annonces marquantes du 17 août 2026.
The "Preparedness" team, dedicated to assessing the most serious risks of border models, was dismantled at the end of July and its missions redistributed to other services.