Auteur/autrice : noflux (Page 2 of 635)

Valais: l’École Ardévaz utilise l’IA pour corriger les examens

12 octobre 2026 /

Depuis la rentrée, les enseignants de l’École Ardévaz, établissement privé de Sion, ne corrigent plus eux-mêmes les examens. Cette tâche est confiée à «Profassist», une plateforme d’intelligence artificielle développée pour aider les enseignants. «C’est une impulsion venue des étudiants, pour qui l’utilisation de l’IA était une évidence», nous explique Alexandre Moulin, le directeur général de l’école, qui souligne que l’outil a été testé en juin et approuvé par les parents, les enseignants et les élèves.

Source: Valais: l’École Ardévaz utilise l’IA pour corriger les examens – 20 minutes

Kenyans Made a Living Writing College Essays. Then A.I. Arrived.

12 octobre 2026 /

His first assignment was a two-page essay on the health benefits of the Garcinia cambogia. It was a fruit he had never heard of, and an essay written for a student he didn’t know anything about. He earned $7 for three hours of work.“That was a lifeline for me, I needed the money,” Mr. Bundi, now 34, said in an interview in Nairobi. He added: “I could not sleep that night. I was ready to burn the midnight oil.”Mr. Bundi estimates he wrote more than 2,500 essays over the next 12 years, sometimes three a day. He balanced completing his public health degree with helping faraway students studying engineering, medicine and computer science to finish theirs. Some students gave him their school login information so he could keep up with their assignments.

Mr. Bundi didn’t know it then, but he was in the boom era of internet gig work. Newly wired to the global internet and hungry for jobs, well-educated youths from places like Kenya, the Philippines and India found opportunities in the crevices of the digital economy. They could make money online doing unglamorous tasks that others further up the economic ladder did not want to do.Then A.I arrived, and the good times ended.

Source: Kenyans Made a Living Writing College Essays. Then A.I. Arrived. – The New York Times

Southern District of New York | North Carolina Man Sentenced To 18 Months In Prison For Super Intelligence-Assisted Music Streaming Fraud | United States Department of Justice

8 octobre 2026 /

SMITH’s hundreds of thousands of AI-generated songs were streamed by his Bot Accounts billions of times.  For example, in April 2023, the entire catalogue of Taylor Swift received 9.3 million streams on YouTube Music from family plan streams, while in the same month, Smith’s Bot Accounts used family plans to fraudulently stream his AI-generated music 80.9 million times.  Smith’s Bot Account streams allowed him to fraudulently obtain more than $8 million in royalties from the Streaming Platforms.

Source: Southern District of New York | North Carolina Man Sentenced To 18 Months In Prison For Super Intelligence-Assisted Music Streaming Fraud | United States Department of Justice

Quand Pangram, l’outil qui détecte l’utilisation des IA, fait irruption dans la vie des éditeurs

30 septembre 2026 /

Il reste à savoir si l’IA, qui puise dans un immense fonds commun de pensée humaine, pourrait susciter l’adhésion des lecteurs justement parce qu’elle reflète ce fonds-là plutôt qu’une pensée singulière. Et si, dans un second temps, le style propre à la machine pourrait devenir familier au point qu’une écriture 100 % humaine finisse par désorienter les lecteurs. La question dépasse largement le monde de l’édition, mais c’est ce secteur qui se trouve en première ligne et, avec lui, la notion d’auteur telle qu’elle existe depuis le XVIIIe siècle.

Source: Quand Pangram, l’outil qui détecte l’utilisation des IA, fait irruption dans la vie des éditeurs

Pangram sait-il vraiment détecter les textes IA ? Notre test sur plus de 1 000 articles – Next

30 septembre 2026 /

Première conclusion. Avec notre corpus d’articles Next, Pangram fait quasiment un sans-faute pour identifier les textes humains (99,8 % de détections correctes) et reste très bon pour identifier les textes générés par IA, avec 95 % de réussite, c’est largement mieux que les 75 détecteurs de notre précédente analyse.

Source: Pangram sait-il vraiment détecter les textes IA ? Notre test sur plus de 1 000 articles – Next

L’IA signe-t-elle la fin de la littérature ?

30 septembre 2026 /

À partir de cette distinction entre impératif de sincérité et exigence formelle, William Marx explique que, dans le cas du récit de Thélyson Orélien, c’est « le miracle de quelqu’un qui, partant d’une condition extrêmement difficile de victime, arrive à trouver la langue, le style qui lui permet de l’exprimer » qui s’effondre. « Quand le style s’écroule, c’est la sincérité elle-même qui disparaît, parce que la sincérité ne reposait que sur cette voix personnelle que l’on croyait entendre dans cette œuvre, et ce scandale n’arrive pas à n’importe quel moment de l’histoire de la littérature« .

Source: L’IA signe-t-elle la fin de la littérature ? | France Culture

Donald Trump rejects calls from tech bosses for AI slowdown

14 septembre 2026 /

The heads of some of the leading US AI groups — Dario Amodei of Anthropic, Sam Altman of OpenAI and SpaceX’s Elon Musk — this weekend called for a more cautious pace of development following dramatic safety warnings over rapidly advancing model capabilities.

But the president said on Sunday that while some “guardrails” were possible, it was imperative for the US to retain its international leadership in the area.

“Look, we’re leading China in AI . . . and, frankly, I want to keep it that way, because whoever wins AI, wins,” Trump told reporters during a trip to his Doonbeg golf course in Ireland. 

“We can put guardrails, we can do this and that, but I think you have a lot of negative forces that are bringing it up that shouldn’t be bringing it up and they’re bringing up things that won’t happen,” he added.

Source: Donald Trump rejects calls from tech bosses for AI slowdown

Industry Leaders Join Open Secure AI Alliance for AI Safety and Security | NVIDIA Blog

30 juillet 2026 /

The recent Hugging Face security incident delivered a clear reminder: cyber defenders need open, frontier agentic systems for self-defense. When closed AI tools — unable to distinguish attackers from defenders — blocked essential forensic analysis, Hugging Face ran the open-weight GLM 5.2 model on its own infrastructure to analyze more than 17,000 actions and contain the intrusion.  

That incident showed a practical truth: when defenders cannot inspect, adapt and run advanced AI on their own infrastructure, their ability to respond is constrained at exactly the moment speed matters most. Companies and countries need open frontier defensive tools and techniques so critical industries can build security systems across a multi-vendor ecosystem and avoid single points of failure.

That is the mission of the Open Secure AI Alliance: to ensure defenders everywhere have open, frontier tools they can trust and control.

Source: Industry Leaders Join Open Secure AI Alliance for AI Safety and Security | NVIDIA Blog

NHS England reprimanded over Palantir data access omission

28 juillet 2026 /

Tom Hegarty, head of communications at the legal campaign group Foxglove, said: “As [the national data guardian] has said, sharing raw patient data with Palantir, after previously promising that wouldn’t happen, has blown a hole through the core principles on health data that are the foundations of trust and consent between the NHS and the public.”

Source: NHS England reprimanded over Palantir data access omission

FLUX 3 – Real World Models: Towards Multimodal Flow Models as the Backbone of Visual Intelligence. | Black Forest Labs

25 juillet 2026 /

FLUX 3 is our new multimodal foundation model. It jointly learns from images, videos, and audio within a unified architecture, because what it needs to learn is not any one of these elements in isolation. Instead, a model must learn a representation of the world: how objects hold together, how things move, and how events sound.

No single modality provides a complete description. Each is a projection of the same underlying reality, captured by different sensors, each of which loses some information in the process. Images capture spatial structures and relationships at a specific point in time. Videos restore the dimension of time and reveal temporal dynamics and physical laws. Audio reveals causal relationships between mechanical phenomena and acoustics that vision alone cannot detect. Language links these perceptions to goals, abstractions, and instructions.

Learn from one and you get a good model of that projection. Learn from all of them at once and their mutual constraints tell you more: the sound has to match the impact, the motion has to obey the mass, the future has to follow from the past. The modalities stop being separate and start being evidence about one underlying reality.

Source: FLUX 3 – Real World Models: Towards Multimodal Flow Models as the Backbone of Visual Intelligence. | Black Forest Labs

« Older posts Newer posts »

Archives

Catégories

© 2026 no-Flux

Theme by Anders Noren — Up ↑