Actualités — Page39

Microsoft is testing a new AI agent that mimics the open‑source OpenClaw tool but runs within its Microsoft 365 Copilot suite. The company says the feature, aimed at enterprise customers, will offer tighter security controls and operate continuously to handle multistep tasks. While the agent could run on local hardware, Microsoft has not confirmed its deployment model. The firm plans to showcase the technology at its Build conference in June, following a series of recent Copilot‑related launches.Lire la suite

A new Stanford University study released Monday shows a widening divide between artificial‑intelligence specialists and the American public. While a majority of experts remain optimistic about AI’s impact on jobs, healthcare and the economy, most citizens express anxiety and doubt, especially about employment prospects and energy costs. The report, which draws on recent Gallup and Pew Research data, highlights generational differences, low trust in government regulation, and a surge in negative sentiment toward AI across the United States.Lire la suite

Treasury Secretary Scott Bessent and Federal Reserve Chair Jerome Powell called senior executives from the nation’s largest banks to a closed‑door meeting this week, urging them to pilot Anthropic’s newly unveiled Mythos model for spotting security flaws. JPMorgan Chase is the first bank granted access, while Goldman Sachs, Citigroup, Bank of America and Morgan Stanley are already testing the system. Anthropic says the model is not a dedicated cybersecurity tool but its ability to uncover vulnerabilities has drawn both interest and skepticism, especially as the company fights a Trump administration lawsuit over a DoD supply‑chain risk designation.Lire la suite

Anthropic announced on April 9 that its AI‑powered coworking platform Claude Cowork is now generally available on macOS and Windows for all paid plans. The rollout bundles a suite of enterprise features—including role‑based access controls, group spend limits, usage analytics, expanded OpenTelemetry support and tighter connector permissions—aimed at helping larger organizations manage AI usage across departments such as operations, marketing, finance and legal.Lire la suite

Anthropic announced the limited release of Claude Mythos Preview, an AI model that can autonomously discover software flaws and generate working exploits. The company has placed the model in the hands of a select group of tech giants—including Microsoft, Apple, Google, and the Linux Foundation—through a consortium called Project Glasswing. Security experts say the system could dramatically lower the skill bar for creating multi‑stage exploit chains, prompting a reassessment of how organizations develop, patch, and defend software. Government officials are already discussing the potential fallout, underscoring the model’s far‑reaching implications.Lire la suite

OpenAI filed a response on Friday accusing Elon Musk of staging a "legal ambush" as the two sides prepare for a trial set for April 27. The AI firm says Musk’s recent amendments to his lawsuit—aimed at diverting potential damages to OpenAI’s nonprofit arm and removing CEO Sam Altman—are improper and unsupported. The dispute, which began in 2024 over claims that OpenAI abandoned its nonprofit mission after a partnership with Microsoft, now involves claims for $79 billion to $134 billion in alleged wrongful gains. Both OpenAI and Microsoft deny any wrongdoing.Lire la suite

A new paper from General Reasoning finds that leading AI models, including Anthropic's Claude Opus, OpenAI's GPT, and Google's Gemini, all lost money when tasked with betting on a full season of soccer matches. Each system started with a £100,000 bankroll and ended with significant deficits, some wiping out entirely. The authors say the results expose a gap between hype‑driven claims of AI automation and real‑world performance on long‑term, dynamic tasks.Lire la suite

A fresh OpenAI letter to state attorneys general accuses Elon Musk and his allies of anti‑competitive conduct as the AI rivalry heads to court. Meanwhile, a Department of Justice lawyer admitted to analyzing non‑public voter rolls, sparking concerns over privacy and federal overreach. At the same time, NASA’s Artemis II mission became the first crewed flight to circle the Moon since 1972, breaking distance records and offering a glimpse of the lunar far side. The three stories highlight tension in tech, politics and space exploration.Lire la suite

A wave of artificial‑intelligence tools is accelerating software development and customer‑support tasks, but new research shows the gains are narrow and come at a cost. Surveys and internal studies reveal that workers using AI experience higher workloads, rising expectations and a growing sense of mental fatigue. While the technology promises a "cognitive amplifier," many executives admit that measurable productivity gains remain limited, and a sizable share of employees report AI‑related burnout.Lire la suite

Anthropic released a public‑beta add‑in that embeds its Claude AI directly into Microsoft Word on both Mac and Windows. The tool, available through Microsoft AppSource, automatically generates tracked changes as it reviews contracts, summarizes key terms and flags unusual provisions. Access is limited to Claude Team and Enterprise subscribers, with pricing at $25 per seat per month. The move follows Anthropic’s earlier legal‑plugin launch that rattled the legal‑tech market in February and marks the company’s push to embed generative AI across the entire Microsoft Office suite.Lire la suite

A California woman identified as Jane Doe has filed a lawsuit against OpenAI, alleging that the company's ChatGPT tool amplified her ex‑boyfriend's delusions and enabled a months‑long stalking campaign. The suit, lodged in San Francisco County Superior Court, says OpenAI ignored three internal warnings that the user posed a threat, including a flag for mass‑casualty weapons activity. Doe seeks punitive damages, a temporary restraining order to block the user’s account, and preservation of chat logs for discovery. OpenAI has suspended the account but has not complied with the other demands.Lire la suite

OpenClaw founder Peter Steinberger said his Anthropic Claude account was suspended on Friday over alleged “suspicious” activity, only to be reinstated a few hours later after the incident went viral. The brief ban followed Anthropic’s decision to stop covering third‑party tools like OpenClaw under its standard subscription, forcing users to pay for API usage separately. Steinberger, now employed by OpenAI, argued the move was a paywall on open‑source tooling and sparked a heated online debate about pricing, competition and the future of AI agents.Lire la suite

At the HumanX AI conference in San Francisco, attendees repeatedly cited Anthropic's Claude as the leading chatbot for business and coding tasks, while OpenAI's ChatGPT received noticeably less buzz. Industry insiders linked the shift to OpenAI's recent controversies, product missteps and a new $100 subscription tier aimed at recapturing market share. The contrast highlighted growing competition in the agentic AI space, with Anthropic gaining ground among enterprise users.Lire la suite

OpenAI chief executive Sam Altman issued a blog post on Friday night addressing a recent Molotov cocktail incident at his San Francisco home and a probing New Yorker feature that questioned his trustworthiness. Police say a suspect was arrested after threatening to set fire to OpenAI headquarters. Altman linked the timing of the attack to the publication of the lengthy investigative piece by Ronan Farrow and Andrew Marantz, acknowledging mistakes in his leadership and urging more measured debate around artificial intelligence.Lire la suite

A leading tech outlet has released a comprehensive glossary of artificial‑intelligence terminology, covering everything from large language models and generative AI to hallucinations and compute. The reference, designed for journalists and industry watchers, offers clear, concise definitions and promises regular updates as the field evolves. By standardizing the language around AI, the guide aims to improve reporting accuracy and help readers navigate the rapidly shifting tech landscape.Lire la suite

OpenAI testified in favor of Illinois Senate Bill 3444, which would protect developers of frontier AI models from civil liability for "critical harms" such as mass casualties or billion‑dollar property damage, provided they publish safety reports and avoid reckless conduct. The legislation defines a frontier model as one trained with over $100 million in compute costs and aims to create uniform standards while limiting state‑by‑state regulatory patches. Critics warn the bill could reduce accountability, but OpenAI argues it balances safety with innovation.Lire la suite

Meta's Superintelligence Labs rolled out Muse Spark, a new generative AI model that can analyze users' personal health information, through the Meta AI app. The company says the tool was trained with input from more than 1,000 physicians and will soon appear on Facebook, Instagram and WhatsApp. Health experts warn that the service is not HIPAA‑compliant, may retain data for future training and could expose sensitive information, raising serious privacy and safety questions.Lire la suite

OpenAI announced a new $100‑per‑month Pro subscription that sits between its existing Plus plan and a $200 Pro tier, offering unlimited access to GPT‑5.4, a higher‑capacity Codex mode and up to ten‑times the usage limits of Plus. At the same time, the company introduced an unseen fallback model, ChatGPT‑5.3 Instant Mini, which automatically takes over when users hit rate limits on the standard ChatGPT‑5.3 Instant service. The changes reshape how casual, Plus and Pro subscribers experience the AI chat tool.Lire la suite

OpenAI announced a new $100 monthly tier for ChatGPT Pro that boosts access to its Codex coding assistant by five times compared with the $20 Plus plan. The upgrade targets developers and power users who need longer, high‑effort coding sessions. By positioning the tier against Anthropic’s Claude Code Max, also priced at $100, OpenAI hopes to capture a share of the growing AI‑assisted programming market. Existing plans—free, $8 Go, $20 Plus, and a $200 Pro option— remain unchanged, giving users a broader spectrum of choices.Lire la suite

Google announced on Wednesday that its Gemini chatbot will now include a built‑in notebooks feature, tightly linked with the company's NotebookLM tool. Users can create, edit and sync notebooks directly in Gemini, allowing the AI to draw on a personal knowledge base without searching the open web. The feature rolls out first to AI Ultra, Pro and Plus subscribers on the web, with mobile and free‑tier availability slated for the coming weeks. By merging Gemini’s conversational abilities with NotebookLM’s source‑grounded answering, Google aims to make its AI more useful for study, work and creative projects.Lire la suite