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
Two weeks after a launch restricted to cybersecurity professionals, Google is making its Edge model available to developers, businesses and Pro and Ultra subscribers.
Following on from a presidential decree issued in September, the scientific funding agency is urging its researchers to replace ‘AI’ with ‘SI’, going beyond what the decree requires.
Researchers hide malicious commands, encoded using the Caesar cipher, within third-party skill files: agents then decrypt them and execute them.
Researchers propose a scale from L0 to L5 to characterise the self-improvement of robots and note that no published work actually reaches the recursive stage.
The French advertising group has exceeded market expectations, confirmed its annual targets and noted that clients are keen to embrace AI in marketing but remain cautious about their major transformation projects.
In *Noema*, the auditor Anna August argues that the registration of models, the prompt reporting of incidents and independent audits would be sufficient to make AI transparent.
A longitudinal study links an Alzheimer’s disease progression score to changes in ‘brain age’ estimated using covariance neural networks.
To find out whether a machine learning model is better at predicting turbulence above large telescopes, a point of comparison is needed: one team proposes two such points.
According to the Institute of Student Employers, 92 per cent of employers have suspected or detected fraud when recruiting graduates, with the undeclared use of AI topping the list.
▶
Cinq actualités en cinq minutes : les modèles Muse Wonder de Meta, AlphaProof Nexus validé par les pairs, l'accès de l'AISI britannique rétabli, les puces Maia de Microsoft et le virage d'Epic Games.
▶
Corey Gallon présente une méthode de relecture en trois niveaux pour repérer les agents IA qui surestiment leur travail, lisent mal les faits ou tirent de mauvaises conclusions.
Rather than predicting the temperature point by point, researchers are compressing atmospheric temperature profiles into latent space – the first step towards more consistent and faster emulators.
Trained on 24 years of satellite data, a ConvLSTM model outperforms simple persistence by up to four years and predicts that soil degradation will continue between now and 2030.
A study compares CLEF and CLEF-Flash using three public breast ultrasound datasets: accuracy varies significantly depending on the prompt and the dataset, with numerous false-positive diagnoses of malignancy.
This three-step learning framework achieves 85.7% simulation success from an on-board vision, and up to 76% of a human's success rate on the real robot.
AI models learn from past observations: where no measurements are taken, they make educated guesses. Down To Earth advocates funding instruments for poorer countries rather than providing them with free forecasts.
The October volume exceeds the previous September record (nearly 1,000 faults). Microsoft credits large language models for discovering these vulnerabilities; Security teams struggle to follow.
The start-up, founded by three German entrepreneurs, aims to connect the AI agents of suppliers’ and customers’ services across different companies to resolve billing disputes and speed up payments.
Following on from Instagram, Meta is allowing Facebook users to specify the topics they want to see more or less of, with recommendations that adjust in real time.
An open letter signed by representatives from 34 countries calls for common safety standards and a register of safety incidents for the most advanced models, arguing that voluntary commitments are insufficient.