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
▶
DW confronte deux juristes, Bonnie Docherty (Harvard, Human Rights Watch) et Laurie R. Blank (Emory) : le droit de la guerre, écrit après 1945, peut-il encadrer des armes guidées par l'IA ?
The judges ruled that it could not be assumed that all buyers had seen the same misleading statement on Tesla’s website: the claimants will have to bring their cases individually.
GPT-6 Astra Pro, followed by teams of Gemini 4 agents, produced plasma equilibria that the physicist Harold Grad had deemed impossible, for less than $3 per problem.
Oracle’s AI campus in New Mexico is set to see its permitted annual greenhouse gas emissions reduced from 8.8 to 5.7 million tonnes; opponents are demanding to see the calculation models.
Kirill Dmitriev, Vladimir Putin’s economic envoy, has spoken of Russian-American data centres powered by small power stations, following a decree by Trump suspending sanctions.
Mayor Daniel Lurie has signed the moratorium passed unanimously by the Board of Supervisors; the planning department is to assess the impact of data centres and propose regulations.
Bitsight tracked a fake ‘penetration testing tool’ across 16 forums, where buyers paste in stolen OpenAI, Anthropic or Google keys: it is the account holders who end up paying for the scans.
Stolen API keys were used to generate charges without customers’ knowledge. OpenCode is calling for them to be revoked and is coordinating a response with Google, OpenAI and Anthropic.
According to openDemocracy, Amazon assigns its sites a score between 100 and 900 based on Reddit posts and internal surveys; Sutton Coldfield and Coventry are at the highest alert level.
The South Korean start-up is now manufacturing its neural processor using Samsung’s 5 nm process, claims to have secured 77 orders from more than ten countries, and is aiming for a stock market listing within three years.
▶
Atelier complet sur l'évaluation d'agents : traces, évaluations par code et par LLM juge, expériences pour valider un correctif, puis passage de Claude Haiku 4.5 à 5.5 preuves à l'appui.
The RemoTAP-Hybrid algorithm, applied to one year’s worth of data from the SPEXone polarimeter, produces the first global seasonal maps of aerosol composition as observed by this instrument.
To replace CFD simulations that are too slow during the sketching phase, two researchers are comparing two ways of teaching a U-Net the geometry of a part.
Based on a single instruction, this agent – which combines an LLM, knowledge of physics and simulations – designs a ring with an emittance of less than 10 pm and recalibrates the model of a real-world facility.
The team responsible for operating the forecasting systems at the LBT and VLT observatories is testing neural networks against the autoregressive method, which is currently the benchmark for one-hour forecasts.
At a hearing in San Jose, Judge Edward Davila ruled in favour of Apple, which accuses OpenAI of having obtained its trade secrets by recruiting hundreds of its employees.
Through its regulatory sandbox, Seoul is exempting IGRIS-A robots from the restrictions imposed on industrial machinery whilst they undergo trials at four factories and construction sites.
A study has scrutinised 21,097 MCP servers and demonstrated how a booby-trapped argument causes a tool to deviate from its stated purpose; 13 security advisories followed, including two CVEs in Anthropic’s Git server.
Every agent – whether an in-house agent, a third party or otherwise – receives a cryptographic token that is verified at runtime; if it strays outside its authorised scope, it may be redirected, restricted or cut off completely.
In the JAMA, researchers from Harvard analysed 2.5 years’ worth of OpenEvidence usage by 950,000 clinicians: more than half of the queries relate to the choice of treatment.