Latest AI News

Published September 23, 2026 at 08:17 Comment 1 𝕏 Print © Geralt - Pixabay. In quick succession, OpenAI and Anthropic have launched new AI systems that are much cheaper than the previous ones. GPT-6 Sol and Luna on one side, Claude Opus 5.5 on the other. The AI price war is intensifying amid controversy over the slowdown of AI. Just minutes apart on Tuesday, OpenAI and Anthropic unveiled new artificial intelligence models. They are all sold for far less th

Published on September 23, 2026 at 10:30 Comment 1 𝕏 Print © Dassault Aviation Two tricolour algorithms have just taken to the skies aboard a fighter jet. Their exact role? We can’t say. Since the abandonment of the European fighter, each French aircraft manufacturer has been charting its own course toward combat AI. Head Music Rafale View the technical specifications The split was finalized on June 8: France and Germany abandoned the joint fighter in the

Published on September 23, 2026 at 3:15 PM Comment 0 𝕏 Print © Unsplash Meta is currently testing a feature that allows AI to make phone calls on your behalf. In some cases, the AI agent could delegate a call to a human. The latter will then make the call based on the instructions the user has provided to the AI. Such a practice inevitably raises privacy concerns. On September 8, Meta officially rolled out its new model, called Muse, in the United States.

Published on September 24, 2026 at 07:35 Comment 0 𝕏 Print © Meta Meta has just unveiled the Muse Charm. This small device, designed to be worn as a pendant, allows you to converse with Muse, the group’s AI assistant, without a smartphone or connected glasses. The gadget will go on sale towards the end of the year, at a price that is still unknown. Mark Zuckerberg had kept the surprise for the end. As the Meta Connect conference was drawing to a close, the

Published on September 24, 2026 at 08:35 Updated on September 24, 2026 at 16:41 Comment 2 𝕏 Print © 01net A version of ChatGPT gained access to an Australian government portal without any authorization. Blocked by a restriction, the AI model looked for ways to get around it. OpenAI took several months to detect the incident and notify the relevant authorities AI systems keep racking up security incidents. Anthony Albanese, the Australian prime minister, re

Published on September 24, 2026 at 3:33 PM Comment 2 𝕏 Print © Montage 01net - Anthropic Hundreds of agents unleashed on viruses for 21 hours, an unprecedented genetic discovery at the end, and a question that remains unanswered: what is it for? The first discovery from Anthropic’s biology lab deserves attention—and scrutiny. Before co-founding Anthropic, Dario Amodei completed a biophysics thesis at Princeton and then worked as a postdoctoral researcher a

Published on September 25, 2026 at 06:33 Comment 1 𝕏 Print © 01net OpenAI was presenting its integration into Siri as the showcase for its partnership with Apple. In a court filing submitted in Texas, the company now describes a feature that almost nobody was using. Apple and OpenAI announced their partnership at the opening of WWDC in June 2024: ChatGPT was set to assist Siri with questions that were too complex for it and with writing, directly from the

Published on September 25, 2026 at 2:00 p.m. Comment 1 𝕏 Print © Google With Live Avatar, Google wants to give Gemini a face. Users can give their avatar a specific appearance (human, animal, animated) and give it a personality. For now, this feature is limited to a specific Google plan. If artificial intelligence is improving day by day, it still badly lacks a face or a “physical” representation. Indeed, the various AI providers are working to fill this g

Published on September 25, 2026 at 4:00 PM Comment 0 𝕏 Print © Google The chips powering Gemini are going to leave Earth. Google is taking its TPUs aboard a test satellite to verify that they survive in space, while a startup backed by Nvidia has already gained a head start. Since the Suncatcher project was announced in November 2025, Google has been entertaining a simple idea: moving some of its artificial intelligence computations into orbit, where the s

Published on September 26, 2026 at 09:11 Updated on September 28, 2026 at 13:02 Comment 5 𝕏 Print © Mariia Shalabaieva OpenAI AI agents searched several U.S. government sites this summer. The Department of Education, the Census Bureau, and the financial watchdog are involved. OpenAI says no data was stolen but acknowledges “misaligned” behavior. Update on September 28 In response to this new misstep, OpenAI decided to suspend the training of its latest AI

Published on September 26, 2026 at 11:15 Comment 6 𝕏 Print © 01net A prize-winning novel in France has been accused of having been written by AI. On the strength of an AI detector presented as virtually infallible, a writer found himself at the heart of a heated controversy. Yet our tests show that the detector can be gamed by a savvy user using a series of tricks. Since September 21, 2026, the Pangram affair has rocked the French literary season. That day

Published on September 28, 2026 at 10:00 Comment 0 𝕏 Print © Unsplash Several iterations of ChatGPT targeted a United Nations data platform. The system recorded more than 16,000 attempted accesses. To achieve their objectives, OpenAI’s artificial intelligences hijacked a game developed by Google. OpenAI’s AIs have slipped up again. This time, a United Nations site ended up in the sights of AI models. Between April and June, autonomous agents bombarded with

Published on September 28, 2026 at 1:00 PM Comment 5 𝕏 Print © Bill Gates - © Kuhlmann / MSC Bill Gates is also raising the alarm about artificial intelligence. If misused, this technology could cause “a billion deaths,” the billionaire estimates. Alongside OpenAI, Anthropic, Google and Elon Musk, the Microsoft co-founder is calling for more stringent government regulation to prevent abuses. Bill Gates does not mince his words about the risks of AI. Interv

Published on September 28, 2026 at 16:02 Comment 0 𝕏 Print © 01net ChatGPT agents published photos sent by users online on their own initiative. OpenAI removed most of the files, without being able to notify the people concerned. OpenAI acknowledged on Friday that its agents had published 53 images from conversations with ChatGPT on public image-hosting sites. The files came from accounts whose users had not disabled the use of their exchanges to train the

Published on September 29, 2026 at 06:31 𝕏 Print © 01net A Cambridge startup would have assembled the largest collection of video game playthroughs ever created. Its target customers: laboratories that teach robots and autonomous vehicles about the real world. Worldmodeldata would have passed the threshold of one million hours of licensed video game gameplay, three months ahead of the target it had set for the end of 2026. Founded last year in Cambridge by

Published on September 29, 2026 at 08:33 Comment 0 𝕏 Print © Anthropic Anthropic has just unveiled Claude Sonnet 5.5, the second AI model in the Claude 5.5 family. Faster and less expensive than its predecessor, it is primarily designed to handle everyday tasks. Anthropic is not letting up on AI announcements. Less than a week after Claude Opus 5.5, the American startup launched Claude Sonnet 5.5, an artificial intelligence that stands out for its speed.Ac

Published on September 29, 2026 at 10:00 Comment 0 𝕏 Print © 01net OpenAI has decided not to launch GPT-6.1 Astra, the next major AI model set to power ChatGPT, scheduled for October. According to the AI giant, it did not meet its safety requirements. The announcement comes amid growing concerns over autonomous AI, which has been involved in several security incidents. OpenAI hits the brakes. The AI giant has abandoned plans to launch GPT-6.1 Astra, its ne

Published on September 29, 2026, at 2:00 p.m. Comment 0 𝕏 Print © GaudiLab Editing a photo with Microsoft’s AI is nothing like a head-to-head contest with the machine. Human reviewers examine the original image and the result, with nothing blurred. Microsoft has humans rate the image edits produced by Copilot in order to train its model. Hundreds of contractors would thus compare the photo submitted by the user with the AI-generated version. The photos rea

Published on September 29, 2026 at 5:45 PM Comment 0 𝕏 Print © OpenAI logo, montage by 01net. Suspended for three weeks, ChatGPT’s Pro subscription is reopening its doors with a surprise in the contract: at the same price, the value of use has halved. The admission comes from OpenAI, just hours before its DevDay. Three weeks ago, OpenAI closed registrations for ChatGPT Pro, its $200-per-month plan, which was overwhelmed by demand. They reopen this Tuesday,

Published on September 29, 2026 at 7:28 PM Comment 1 𝕏 Print © 01net Sam Altman has just unveiled “Dots” at the opening keynote of OpenAI’s Dev Days. These AI agents, integrated into ChatGPT, work around the clock on behalf of the user. The company is also launching “ChatGPT Spaces,” a shared workspace for AI agents. OpenAI wants its AI to stop waiting for commands to act. At the opening keynote of the Dev Days, Sam Altman, the head of OpenAI, unveiled “Do