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
▶
Eli Cohen (Snyk) explique pourquoi un test d'intrusion une ou deux fois par an ne suffit plus face à des attaquants qui enchaînent des failles avec l'IA en quelques minutes.
Trained on annotations from Galaxy Zoo 3D and images from the DESI Legacy Survey, this publicly available tool is designed for the major surveys to come, which manual catalogues can no longer keep up with.
Researchers have added a generative layer to the smaller version of Aurora, which is approaching state-of-the-art performance on aggregate metrics and rivals the larger version.
According to Crunchbase, the total figure is down 35 per cent quarter-on-quarter due to the absence of major funding rounds from OpenAI or Anthropic, but is up 50 per cent year-on-year; the acquisition of Hugging Face dominates the list of deals.
The group says it took action against 33.2 million pieces of child sexual exploitation material in the first half of the year and has introduced an AI agent tasked with bypassing its own safeguards.
The Virtual Biology Initiative is set to produce open data to train models capable of predicting how cells will respond to diseases and treatments.
The NGO believes that the teenage version of ChatGPT fails to alert parents when it should and always does their homework; OpenAI questions the testing method.
Led by Insight Partners alongside a16z and M12, the Series B round brings the total raised by the start-up – which automates the entire ‘order-to-cash’ cycle – to $93 million.
Previously available only to journalists and researchers, the tool recognises the SynthID watermark used by Google, as well as by OpenAI, Nvidia and Kakao.
The first Haiku model to feature a performance adjustment, the new model is designed for high-volume, low-cost tasks; Anthropic is also halving the price of cached queries for Sonnet 5.5.
▶
Rafal Wilinski (Runlayer) montre comment des agents rédigent leurs propres compétences à partir de leurs essais réussis et ratés, et les partagent via MCP.
Rob Bonta is demanding further answers regarding the cybersecurity incidents linked to OpenAI’s models, whilst fifteen states are conducting their own investigations.
Trained on a vast dataset of verified calibrations, the model significantly reduces detection errors and could be used in next-generation radio telescopes.
Led by Bret Taylor, this open standard is designed to enable websites to recognise an agent acting on behalf of a real person and to track what that agent is doing.
Defence, red teaming or specialised access: verified security teams are granted less restricted access. Glasswing claims to have identified over 129,000 confirmed vulnerabilities.
In Muhos and Kajaani, land clearing and earthworks must be halted by 23 October, whilst the environmental impact assessments are being completed.
Led by Robot Ventures alongside Nvidia, M12 (Microsoft) and Samsung, the funding round is set to transform a widely downloaded app into a paid service for businesses.
Used in more than 50 hospitals, its software analyses care staff’s notes using language models to flag up malnutrition or delirium before they go unnoticed.
Its autonomous electric railcars, which run in trains without couplings, aim to bring rail transport back into the picture for journeys of less than 800 km, which are currently dominated by lorries.
Available in the browser to US adults, Playground creates a playable game from a natural language description, without writing a single line of code.