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
False positives, bias against non-English-speaking students and legal proceedings have got the better of automatic detection. Institutions are falling back on oral and proctored exams.
Each satellite must carry 72 Rubin GPUs and 36 Vera processors under 250 kW of peak power. The prototype is expected in early 2027, with a constellation of one million computing nodes on the horizon.
The thirtieth Google employee is leaving with Sanjay Ghemawat, Quoc Le and Oriol Vinyals to set up a public interest company that wants to automate the experimental loop of scientific research.
Rest of World investigates economies where macroeconomic indicators are improving thanks to AI, whilst employment is contracting.
Figure AI’s BotQ factory has gone from producing one robot a day to one robot an hour in less than four months – a 24-fold increase that is set to result in 100,000 units being produced in four years.
A design flaw in the Langflow agent platform allows for remote code execution without authentication. CISA has added it to its list of actively exploited flaws.
A Stanford study published in Nature Human Behaviour crosses questionnaires and real conversation histories: it is not intensive use that harms, but confidential use in people who are already poorly supported.
The British AI Security Institute details actions outside the scope of its cyber tests, including the first documented attempt at social engineering carried out by an agent on his own initiative.
The community college network is launching an undergraduate degree in artificial intelligence and machine learning development, with the first courses starting on August 20, 2026.
Precision is becoming the number one driver of adoption, ahead of reducing arduousness and safety, according to BuiltWorlds' annual barometer.
Published in Genome Medicine, the algorithm cross-references clinical notes and sequencing to classify candidate variants, including to reinterpret files that have remained undiagnosed.
The British institute’s 2026 survey found that 74 per cent of agencies use AI on at least some of their projects, and that one in five has integrated it into several of its processes.
This speech recognition model specialized in medical vocabulary greatly reduces the error rate on the names of drugs, diagnoses and clinical procedures.
The Work AI Index 2026 report paints a picture of a paradox: widespread adoption of AI in businesses, yet an organisational impact deemed disappointing by the vast majority of senior management.
The companies surveyed say they are adjusting their recruitment plans and training their staff rather than cutting jobs, despite a sharp rise in the adoption of AI over the past year.
More and more teachers are using AI to prepare their lessons, tailor their teaching methods or write feedback for pupils, whilst remaining cautious about its use in direct teaching.
▶
Une vidéo qui questionne la notion de « raisonnement » chez les LLM : passer des tests ne prouve pas la compréhension.
A week after its formation, the industry alliance of more than 120 companies presented a mechanism for confidential sharing of AI-related security incidents.
Designed for workspaces managed by educational institutions, these extensions cover primary and secondary education, higher education and students.
Unveiled at the Black Hat conference, this system controls access and permissions for AI agents on a per-tool basis, with the option to reverse their actions.