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
$350 million raised in Series E, led by Insight Partners and S32, driven by an on-demand data business that has grown more than eighteen times in one year.
After three years of negotiations without an agreement, the practitioners of two Allina Health hospitals are demanding a right to review the place of algorithms in care and billing.
Prohibition of training models on student data, no individual monitoring, no decision without human supervision: commitments become contractual from November 1st.
A University of Chicago report supported by the Rockefeller Foundation believes that AI models can finally make forecasting affordable — as long as they don't replicate existing inequalities.
The International Federation of Robotics' World Robotics 2026 report estimates 603,000 installations in 2025, including 354,000 in China, and attributes the rest to AI and machine vision.
On 24 September, the two capitals finally talk about artificial intelligence. The agenda is broad, the ambition deliberately modest: to set definitions, not a framework.
Twelve days after the launch, Amazon blocked Meta's agent at checkout and cited hidden browsing and unauthorized credential processing.
Attorney General Niki Sharma is calling for "a clearer path to human and corporate accountability for the actions of AI" after the investigation into the Tumbler Ridge shooting.
Cisco Talos describes the first documented implant whose tactical control is entrusted to a panel of business models, and publishes CAIRN to track this type of code.
▶
The AI Daily Brief compare les deux salves de modèles publiées à quatre-vingt-dix minutes d'intervalle par Anthropic et OpenAI, et ce que leur baisse de prix simultanée change pour les développeurs.
Two announcements separated by an hour and a half: Anthropic releases Opus 5.5, OpenAI responds with GPT-6 Sol and GPT-6 Luna, billed half as much as the 5.6 series.
▶
Tour d'horizon des annonces de la semaine du 25 septembre 2026 : Meta Connect 2026, la sortie de Claude Opus 5.5 et les évolutions de ChatGPT.
▶
Matt Wolfe passe en revue les annonces IA de la semaine et s'arrête sur les signaux d'un ralentissement du secteur, entre sorties de modèles et inquiétudes de marché.
UK Power Networks has warned Nscale that its Loughton site will not have its power until the beginning or middle of the next decade, compared to 2027 announced.
The boss of Dymocks Education believes that the human tutoring model is no longer competitive with AI subscriptions at $30 per month.
Facility-0 gives models control over 40 real instruments; on the SciUniverse bench, Claude reaches 45.3% compared to 32.5% for GPT-6.
A test bench places between 3 and 20 agents in asymmetric roles within an MMORPG-style sandbox. None of the models tested achieved a success rate of more than 52 per cent, and a new metric measures the effort that is actually useful.
Liang Wenfeng presented these figures to investors: 82.9% gross margin on API until July, and a project to raise about 7.5 billion before a listing in Shanghai.
With dormakaba and LEGIC, the inspection robot now passes through the access-controlled doors like an identified employee, tested for several weeks in Ireland.
Two decrees signed a day apart, in the wake of Gavin Newsom's decree in California: the first sets up a body of experts with the state, the second instructs its IT director to write independent audit standards.