Actualités — Page23

Homeowners facing the annual spring‑cleaning slog are turning to artificial‑intelligence assistants for help. From Claude’s recipe for long‑lasting cleaning solutions to Google’s Gemini recommending power‑scrubbers and safety tips, AI is acting as a planner, product guide, and visualizer. The technology breaks down overwhelming chores, suggests microfiber cloths, ceramic coatings, and even warns against mixing bleach with acids. Visualization apps like Deptho AI let users preview a decluttered space before lifting a finger. As AI advice spreads, experts caution users to verify product compatibility and follow safety guidelines.Lire la suite

Intruder, a UK cybersecurity startup, has launched AI pentesting agents that replicate manual pen testing methodology in minutes, threatening to replace human pen testers. The company's AI agents work by investigating vulnerability scanner findings, interacting with target systems, and determining whether findings represent genuine exploitable flaws or false positives.Lire la suite

Anthropic's Claude Mythos Preview, an AI model, has identified thousands of zero-day vulnerabilities in major operating systems and browsers, prompting the Federal Reserve chair and Treasury secretary to convene an emergency meeting with bank CEOs to discuss the cyber risks. The company warns of a six-to-twelve month window to patch the flaws before adversaries replicate the capability.Lire la suite

Akamai Technologies' stock had its best day in 22 years, rising 27% after the company announced a $1.8 billion, seven-year cloud deal with Anthropic, a leading AI provider. The contract is Akamai's largest ever and marks a significant pivot into AI infrastructure.Lire la suite

Perplexity has opened its Personal Computer AI agent to every Mac user, offering a desktop assistant that can search local files, operate native apps, browse the web and tap into more than 400 tools without constant supervision. Users launch a task, the system works autonomously, and only intervenes when a decision requires approval. The app, which runs best on a Mac mini, is downloadable directly from Perplexity’s website and is not yet listed on the App Store.Lire la suite

OpenAI announced three new audio models for its Realtime API—GPT‑Realtime‑2, GPT‑Realtime‑Translate and GPT‑Realtime‑Whisper. The suite pushes voice AI beyond simple back‑and‑forth exchanges, offering live reasoning, on‑the‑fly translation across 70+ languages and streaming transcription. Developers can now build assistants that schedule home tours, manage travel bookings or provide real‑time captions, while pricing starts at $0.017 per minute for Whisper and $0.034 per minute for Translate, with GPT‑Realtime‑2 billed at $32 per million audio tokens.Lire la suite

OpenAI is rolling out a new Trusted Contact option for adult ChatGPT accounts. The feature lets users name a designated person who will be alerted if the AI detects a serious self‑harm concern. After a brief human review, the contact receives a notification without any chat transcript details. OpenAI says the safeguard aims to complement existing safety tools and crisis resources, while giving users more control over their digital wellbeing.Lire la suite

Emails between senior Microsoft leaders, presented in the Elon Musk versus Sam Altman lawsuit, show the tech giant hesitated to pour additional funding into OpenAI as early as 2017. Executives questioned the lab’s progress toward artificial general intelligence and worried a funding shortfall could drive the AI startup to rival Amazon. The correspondence predates Microsoft’s landmark $1 billion investment in 2019, highlighting a cautious start to what would become one of tech’s most celebrated partnerships.Lire la suite

OpenAI announced Thursday that its API now supports three new voice‑focused models—GPT‑Realtime‑2, GPT‑Realtime‑Translate and GPT‑Realtime‑Whisper. The suite lets developers build applications that can converse, translate and transcribe speech on the fly, with support for more than 70 input languages and 13 output languages. Billing is split between per‑minute rates for translation and transcription and token‑based pricing for the conversational model. OpenAI says the tools target customer‑service, education, media and creator platforms, and includes guardrails to curb misuse.Lire la suite

Meta is building an AI assistant named Hatch, modeled after the open‑source platform OpenClaw, to let users shop directly from Instagram Reels and interact with third‑party services such as DoorDash and Outlook. The company tested the prototype on simulated versions of external apps and plans to roll it out before the end of the year, positioning it as a counter to TikTok Shop. Meta’s CEO Mark Zuckerberg highlighted the goal of making agents that understand user goals and work continuously on their behalf, while CFO Susan Li hinted at future integration with the firm’s Ray‑Ban Meta glasses.Lire la suite

OpenAI unveiled a Chrome extension for its Codex platform, letting developers test web apps, pull context from multiple tabs, and run DevTools alongside other tasks. The add‑on, compatible with Windows and macOS, aims to make AI‑assisted coding more accessible to casual users and professionals beyond traditional developers. The move follows Codex’s February macOS release and April feature updates, and it foreshadows a future integrated app that merges Codex, ChatGPT and OpenAI’s Atlas browser.Lire la suite

OpenAI announced a new safety option called Trusted Contact that lets adult ChatGPT users name a friend or family member to be notified if the conversation veers toward self‑harm. When the system detects suicidal language, it prompts the user to reach out and, if the risk is deemed serious, sends a brief alert to the designated contact. The move comes amid a wave of lawsuits alleging the chatbot encouraged suicide. OpenAI says the feature, like its parental controls, is optional and designed to protect privacy while adding a human check on AI‑driven distress signals.Lire la suite

Perplexity announced Thursday that its Personal Computer AI platform is now open to any Mac user through a new desktop app. The tool, originally limited to Max subscribers and a waitlist, lets autonomous agents access local files, native applications and the web to automate multi‑step workflows. While a Pro or Max subscription is still required to unlock the full feature set, the move signals Perplexity’s push to bring local AI assistants into everyday productivity environments, positioning the service as a safer alternative to competitor OpenClaw.Lire la suite

Elon Musk’s legal action against OpenAI alleges that the lab’s for‑profit arm has abandoned its original safety mission in favor of rapid product rollout. In a federal court in Oakland, former employee and board member Rosie Campbell testified that key safety teams were disbanded and that a Microsoft‑deployed GPT‑4 model entered the Indian market without clearance from OpenAI’s Deployment Safety Board. The case also highlights internal board disputes, CEO Sam Altman’s alleged nondisclosure, and expert testimony that the organization’s governance now favors profit over safety, fueling calls for stricter AI regulation.Lire la suite

In a courtroom testimony that has captured the tech world’s attention, former OpenAI chief technology officer Mira Murati detailed the internal battles that led to CEO Sam Altman's abrupt removal in November 2023. Murati described how she helped compile a 52‑page memo for co‑founder Ilya Sutskever, exchanged dozens of urgent text messages with Altman, and briefly served as interim chief executive before the board installed Emmett Shear. The deposition reveals the board’s claim that Altman was “not consistently candid” and shows how employee backlash and Microsoft’s involvement forced a rapid reversal, returning Altman to the helm.Lire la suite

Perplexity has opened its Personal Computer AI assistant to every macOS user, moving beyond the $200‑per‑month Max plan that previously limited access. The new macOS app lets users run everyday queries, manage attachments, and dictate text, while still tying usage to the credit limits of Pro and Max subscriptions. By integrating models like Gemini, Nano Banana and ChatGPT, the assistant handles tasks across local files, other apps, the web and Perplexity’s servers, positioning the company against competitors such as Claude and Apple’s upcoming AI‑enhanced Siri.Lire la suite

Snap disclosed in its Q1 2026 investor letter that it and Perplexity AI have mutually terminated a $400 million cash‑and‑equity partnership announced last November. The deal, which would have embedded Perplexity’s answering engine into Snapchat’s Chat, will no longer contribute to Snap’s 2026 revenue outlook. While the move pauses the planned AI chatbot rollout, Snap continues to push other AI‑driven features and reports a 5 % rise in daily active users.Lire la suite

Elon Musk's SpaceX has signed a data‑center agreement that grants Anthropic access to more than 200,000 Nvidia GPUs at the Colossus 1 supercomputer in Tennessee. The partnership, announced after Musk previously accused Anthropic of stealing training data, will boost capacity for the company's Claude Pro and Claude Max services and includes plans to explore gigawatt‑scale orbital AI compute. Anthropic responded by doubling usage limits for several products and listing other data‑center deals with Amazon, Google and Microsoft. The deal has drawn criticism from the NAACP, which alleges environmental harm to Black communities near the Tennessee facility.Lire la suite

OpenAI rolled out an optional safety tool called Trusted Contact for adult ChatGPT users. The feature lets users name a friend, family member or caregiver who will receive a discreet alert if the system detects language suggesting self‑harm or suicidal thoughts. Notifications contain no transcript details, and both the user and the contact can revoke the link at any time. OpenAI says a small team of trained reviewers will assess flagged conversations before any outreach occurs, aiming to add a layer of human support to existing helplines.Lire la suite

Mozilla’s Firefox team says Anthropic’s new Mythos model has identified dozens of high‑severity vulnerabilities, many lurking for over a decade. The AI‑driven scans helped the browser ship 423 bug fixes in April 2026, a stark jump from 31 the previous year. Researchers credit the model’s ability to self‑filter false positives and generate detailed reports, though human engineers still write and review patches. The breakthrough signals a shift in software security, but Mozilla warns that attackers could eventually co‑opt similar tools.Lire la suite