Microsoft is starting to combine its consumer and commercial Copilot assistants into a single app experience. According to The Verge, users of the Copilot and Microsoft 365 Copilot apps will be moved into a unified Microsoft Copilot app that supports both personal and work accounts and uses updated branding.
Writer has released Palmyra X6, a new flagship model for enterprise agent workloads, alongside a rebuilt orchestration harness and new governance tools aimed at controlling token usage and operating costs. The company says the updated stack cuts average agent costs by 52%, improves speed by 48%, and raises quality by 10% when paired with Palmyra X6. Writer also says Palmyra X6 is a post-trained version of Z.ai's open-weight GLM-5.2 model running on U.S. infrastructure.
Ryanair has added Google Cloud to its infrastructure plans while keeping Amazon Web Services in place, according to The Register. The airline said it will use Google Gemini and related cloud services for operational workloads including crew management, forecasting, and maintenance, framing the move as a dual-cloud strategy rather than a replacement of AWS.
Meta is donating 15,000 Ray-Ban Meta smart glasses to Vision Ireland for blind and visually impaired people in Ireland, according to Engadget. The program centers on accessibility use cases for the AI-enabled glasses, including hands-free assistance for people with sight loss.
The Wall Street Journal reports that Google DeepMind CEO Demis Hassabis discussed creating a new AI oversight entity with leaders of other AI labs and with Trump administration officials, including Treasury Secretary Scott Bessent. The report describes a policy proposal under consideration ahead of a broader government shake-up, rather than a finalized rule or formally launched institution.
The Register reports that OpenAI's ad service may continue billing customers for up to 24 hours after they pause a campaign. The article describes this as a product and billing-policy issue affecting advertisers using OpenAI's ad platform rather than a broader company or model announcement.
The Wall Street Journal reports that Apple is in talks with publishers about paying to license news and other content for use in Siri. The reported discussions suggest Apple is seeking more current, authoritative information sources to improve the quality and timeliness of its AI-powered assistant responses.
Anthropic has added a Chrome browser extension that lets users run Claude Cowork in a sidebar while they browse the web. The update expands access to Anthropic's agent product beyond its earlier smartphone rollout and gives users a persistent in-browser surface for ongoing conversations and task management.
Ars Technica reports on ShieldFont, a newly introduced font intended to make scraped webpage text unreliable for AI systems while keeping pages legible to human readers. The tool is presented as an anti-scraping defense aimed at disrupting training-data collection from the open web rather than blocking access outright.
The Register reports that a cyberattack against Taiwan’s nuclear safety agency involved "near-autonomous" AI agents or agentic tooling. Based on the article framing provided, the incident is being presented as a discrete security event in which AI-assisted offensive techniques were used against a sensitive government target rather than as a product announcement or research result.
Ars Technica reports that attackers compromised an AI-related software package and used it to scrape and exfiltrate terabytes of credentials from about 2,500 users. Based on the report, the incident is a software supply-chain attack with downstream exposure affecting users of the compromised package rather than a breach limited to a single AI company account database.
Futurism reports that Miles Brundage, described as OpenAI’s head of ethics or a senior ethics and policy figure, has left the company. If confirmed, the departure would mark another governance-related personnel change at OpenAI, though the article as provided offers limited detail about the circumstances and scope of the role.
Twitch has introduced a setting that lets streamers opt out of having their streams, VODs, clips, chats, and channel text or images used in future training of Amazon generative AI models. Ars Technica reports the platform had allowed such content to be used for Amazon AI development for years under existing terms, while the new control applies prospectively rather than undoing past use.
xAI has released Grok 4.6, a new frontier model positioned for long-running agents, coding, and knowledge-work tasks. VentureBeat reports the model reached 61 on Artificial Analysis's Intelligence Index, surpassing Moonshot AI's Kimi K3 and tying OpenAI's GPT-5.6 Sol Max, while keeping API pricing at $2 per million input tokens and $6 per million output tokens.
Automattic has launched an Android version of Mesh, its AI-powered contacts and relationship management app. The release expands Mesh beyond its earlier availability and gives Android users native access to the company's consumer CRM-style product.
Sen. Bernie Sanders publicly warned AI company leaders to stop building machines that humans cannot control, according to Futurism. The article describes the statement as a dated intervention in the public debate over AI safety and corporate responsibility rather than a formal legislative or regulatory action.
Facebook has officially rolled out a standalone Creator Studio app that includes an AI creator assistant. According to TechCrunch, the app offers personalized recommendations based on a creator's content style, performance, audience engagement, and stated goals.
Google has revealed its 2026 consumer hardware lineup, including the Pixel 11 phones, Pixel Watch 5, and Pixel Tag. Coverage from Ars Technica says the launch emphasized AI features and curated user experiences more than major specification upgrades across the devices. This broader lineup announcement subsumes the previously tracked Pixel Watch 5 launch as part of a single hardware event.
Futurism reports that an internal Google document indicates an AI system used in hiring or recruiting may have automatically rejected or filtered out some applicants who were qualified for roles. Based on the article framing provided, the report centers on an alleged internal failure in Google’s HR screening workflow rather than a consumer product launch or policy action.
Futurism reports that a farmer says advice generated by an AI system contributed to the loss of 25 acres of sesame crop. The article frames the incident as a real-world failure case in which erroneous AI guidance was allegedly followed in an agricultural setting with costly consequences.
