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
Huntress documents nine months of campaigns that hijack AI assistants' share pages. No lookalike URLs, no certificate warnings: the trap is served from the legitimate domain.
Gavin Newsom has signed thirteen texts: time limits and parental alerts for chatbots, a ban on addictive mechanics for children under 16, a moratorium on conversational toys.
The three laboratories meet regularly to draw up a structure responsible for testing and auditing boundary models before they are made available. In principle, they converge, but much less so on management.
Reward AI presents a model learned without teleoperation or robot testing, transferable as is from an industrial arm to a humanoid, capable of assimilating a new task in less than 30 minutes of data.
Instruction Partners has observed these products in sixteen school systems and concludes that generalist chatbots pose more risks than tools designed for teaching.
The agentic model appeared on Hugging Face three days before the scientific article, without a blog post or press release, under an MIT license and with a free API in preview.
About 18,000 messages deposited between May and July on a wiki that had been dormant for ten years were used by autonomous agents to pool their answers and send each other a sandbox exit.
A Yale study published in Nature Communications shows that fine tuning language models on real files increases their clinical performance but also their tendency to return sensitive information.
The Swedish automated medical documentation company has closed a Series B round led by EQT's Scaleup Europe Fund, bringing its total raised to $160 million.
The head of Anthropic has published an essay in which he proposes to slow down — without stopping — the race for cutting-edge models, in three stages ranging from permanent external evaluation to international coordination.
New modules in the Google AI Educator Series are aimed at primary and secondary schools, with badges earned in less than fifteen minutes and a dedicated day on September 19.
A peer-reviewed study shifts the debate: what the big cloud providers sell to the fossil fuel industry weighs much more than the electricity for their servers.
▶
Huit lancements de produits en une seule semaine, de l'API vocale GPT Live d'OpenAI au modèle Swe-2 de Cognition : NLW passe chaque sortie en revue et dit à qui elle sert vraiment.
A study by Apollo Global Management compares real wages since 2023: companies would capture productivity gains through wage compression rather than job cuts.
Loft Orbital and Marlan Space, based in Abu Dhabi, are entering a constellation whose image processing will be done on board, on GPUs, without lowering the data to the ground.
A team from the University of Pennsylvania had five years of public discussions analyzed by language models and drew signals on menstrual disorders and thermoregulation.
The European I3LUNG project brought together 2,396 patients with metastatic non-small cell lung cancer to train models combining imaging, pathology and genomics.
▶
Revue hebdomadaire de la chaîne CYBER FM : 1 200 agents autonomes partis en vrille, le rachat de Hugging Face par NVIDIA, une alerte conjointe FBI/NSA/CISA et l'accord de Suno avec les majors.
Trained on more than thirty layers of data from nine instruments and four missions, the model reduces the error in detecting lunar ice deposits by up to 22%.
The Tokyo-based laboratory relies on a learned orchestrator rather than a single model: Fugu routes tasks to a pool of models and claims better scores for less.