Topic: #microsoft

Microsoft is launching a revamped Microsoft 365 Copilot that loads roughly twice as fast and presents responses in a cleaner, more scannable layout. A new progressive disclosure feature surfaces tools and controls based on your prompt instead of dumping all options up front. The upgraded prompt box also supports inline text formatting and expands to fit longer inputs.

Microsoft and Anthropic CEOs are courting Canberra, looking for stable bases to train massive AI models that need land, renewable energy and political stability — resources the US can no longer easily provide. Peter Lewis argues Australia should not just become a server farm: an AI sovereign wealth fund could let citizens share in the profits of the looming gold rush.

Microsoft Research introduces Vega, a new approach for proving digital identity using zero-knowledge proofs. Instead of sharing entire credentials, only a single cryptographic proof is transmitted for exactly the claim being verified. Performance is designed to be production-ready and work inside real apps.

Microsoft Copilot is taking over the Excel tasks where many users get stuck: formulas, clean charts, and dashboards. David Fortin demonstrates how Copilot can turn a raw dataset into a complete sales dashboard with product, regional, and dynamic filter charts. Most manual formatting goes away, freeing users to focus on analysis.

Microsoft is rolling out nine updates to its M365 Copilot suite aimed at smoother workflows and higher productivity. The headline change is a redesigned app launcher (the "waffle") that lets users pin frequently used apps and reach the new "Create" module for images and videos in fewer clicks. Search, agent and integration features are expanded too.

A Verge writer notes that Gemini keeps creeping into more and more Google apps — from Gmail and Drive to little sparkle icons across the suite. What started as subtle nudges now feels like a relentless "Gemini everywhere" push. The obvious template is Microsoft, which spammed Copilot shortcuts across Windows 11 to widespread user irritation.

Microsoft Research has posted follow-up notes to its paper LLMs Corrupt Your Documents When You Delegate. The researchers clarify what the study actually shows and what it does not: AI agents in delegated workflows do not always stay clean and can quietly alter documents over time.

Lawyers for Elon Musk and OpenAI delivered their closing arguments Thursday in the high-profile federal trial. The nine-person jury begins deliberations next week, with the outcome hinging on whether Sam Altman and OpenAI broke earlier agreements by restructuring the lab into a for-profit. A verdict could carry broad implications for governance at other AI labs.

Microsoft began rolling out Claude Code to thousands of its developers, designers, and project managers in December. Internal sources say Anthropic's tool became extremely popular — perhaps too popular. Now Microsoft is pulling most of those licenses and pushing its devs toward Copilot CLI instead.

Microsoft researchers have flagged a problem that may unsettle companies betting big on workplace AI: employees who hand routine decisions to language models can lose the habit of critical thinking. The finding sits awkwardly with Microsoft's own push to roll out AI tools across knowledge work.

Karen Hao argues that fixating on whether Altman is untrustworthy — or whether Musk is even less so — distracts from the deeper AI problem. The two ex-OpenAI co-founders are now in a vicious courtroom feud in California: Musk alleges Altman and OpenAI president Greg Brockman tricked him into funding the nonprofit before restructuring it into a for-profit entity.

Microsoft Edge is rolling out a Copilot update that lets the AI chatbot read across all your open tabs. You can ask it to compare products, summarize articles, or answer questions about tab content, and Microsoft lets you toggle which experiences are on. The company is also retiring Copilot Mode, which had similar tab-reading abilities plus agentic features like booking reservations on your behalf.

Introducing GridSFM, a small foundation model that can predict AC optimal power flow in milliseconds, boosting efficiency and unlocking cost savings. Learn how GridSFM gives grid operators direct visibility into congestion, stability, and system health. The post GridSFM: A new, small foundation model for the electric grid appeared first on Microsoft Research.

Microsoft Research expands MatterSim with faster large-scale simulations and a new multi-task model called MatterSim-MT, which predicts properties beyond potential energy surfaces alone. Conductivity, stability and more come from a single model. A meaningful step in both throughput and scope for AI-driven materials science.

Microsoft's Agent 365 is a centralized platform for managing AI agents, hooking into Microsoft Purview, Entra and Defender. Per Microsoft Mechanics, the focus is least-privilege access so agents only get the rights they truly need. Geeky Gadgets walks through how teams can enforce security and compliance centrally.

Microsoft 365 Copilot introduces AI-driven capabilities designed to support a wide range of professional tasks. In his guide, Mike Tholfsen highlights five practical strategies, starting with the Mode Switcher, which adjusts Copilot’s responses based on your specific requirements.

Court documents from the ongoing Musk v. Altman trial reveal rare communications between Microsoft's top executives about investing in OpenAI and fears the AI startup could 'storm off to Amazon' and 'shit-talk' Microsoft. Days after OpenAI's bot beat a Dota 2 pro in 2017, Sam Altman responded to Satya Nadella's congratulations email with a proposal for a far bigger partnership to fund OpenAI's next AI research phase.

Microsoft has introduced the OpenAI Image Gen 2 model into its Copilot platform, offering users a streamlined way to create high-quality visuals directly within Microsoft 365. As highlighted by Mike Tholfsen, this integration supports diverse needs, from designing infographics for business presentations to generating educational materials like comic strips or interactive diagrams. A key feature […] The post What OpenAI’s New Image Gen 2 Model Means for…

Bargains are disappearing and the cost of gadgets such as MacBooks and PS5s is rising as AI competes for memory chips The end of the cheap laptop, the bargain phone and affordable games consoles may be on the horizon. Not because new models are more hi-tech, but because the cost of computer components has shot up.

Microsoft, Google DeepMind and xAI products to be vetted for cybersecurity, biosecurity and chemical weapons risks The US government has struck deals with Google DeepMind, Microsoft and xAI to review early versions of their new AI models before they are released to the public. The Center for AI Standards and Innovation (CAISI), part of the US Department of Commerce, announced the agreements on Tuesday, saying the review process would be key to understan…

Microsoft researchers share advances in building and operating large-scale distributed systems, spanning datacenters, networking, and the growing intersection with AI during NSDI ’26. The post Microsoft at NSDI 2026: Advances in large-scale networked systems appeared first on Microsoft Research.

The Pentagon has signed agreements with leading AI firms, including Microsoft, Amazon and Google, advancing military capabilities amid an ongoing dispute over safeguards. The deal marks another step toward tighter integration of commercial AI research with military use cases.

The Pentagon has struck deals with OpenAI, Google, Microsoft, Amazon, Nvidia, Elon Musk's xAI, and the startup Reflection, allowing the agency to use their AI tools in classified settings. The Defense Department has left out Anthropic — which it previously used for classified information — after declaring it a supply-chain risk.

Microsoft is launching a new AI agent inside Word, designed specifically for legal teams. The Legal Agent handles document edits, negotiation history, and complex contracts.

Safe agents don’t guarantee a safe ecosystem of interconnected agents. Microsoft Research examines what breaks when AI agents interact and why network-level risks require new approaches. The post Red-teaming a network of agents: Understanding what breaks when AI agents interact at scale appeared first on Microsoft Research.

Google, Amazon, Microsoft and Meta reported more than $130 billion in quarterly capital expenditures on Wednesday, primarily for building AI data centers. And there's more to come: all four companies have signaled even larger investments in upcoming quarters, pushing hyperscaler capex into an entirely new tier.

If you want to gauge whether the AI bubble is bursting, Oracle is the publicly traded company to watch. The database giant has burned its boats and pivoted entirely to AI — but not as a foundation model builder like OpenAI, nor as a pure neocloud like CoreWeave. Instead, Oracle is making an audacious bet as a software-as-a-service company on a very specific version of the AI future.

Microsoft Teams introduces Copilot call delegation, an AI feature that helps professionals prioritize their most important calls. It filters spam, surfaces caller context, and schedules follow-ups automatically. Through integration with Microsoft Bookings, callers can even book meetings directly — no human gatekeeping required.

OpenAI's revised Microsoft pact lets it sell AI models across multiple clouds, enabling a likely expansion with Amazon and broader enterprise distribution. Why it matters: The shift ends OpenAI's effective cloud exclusivity, widening its reach to customers using AWS, Google Cloud or others — and intensifying AI platform competition.

One of the most popular Linux distributions is about to get an influx of AI features. Canonical VP of engineering Jon Seager shared a blog post detailing plans to add AI features to Ubuntu over the next year. The features will come in two forms: as a means of enhancing existing OS functionality with AI models in the background, and as 'AI native' features and workflows for those who want them.

OpenAI and Microsoft's partnership-turned-situationship just got even less committed. A clause about artificial general intelligence, which has for years dictated the future of their deal, has officially been dropped.

Anthropic is expanding Claude's connected services to include lifestyle apps. Users can now link their accounts with AllTrails, Audible, Booking. com, Instacart, Spotify, TripAdvisor, Uber, and more.

Meta and Microsoft are cutting thousands of employees while simultaneously investing heavily in AI. Meta will lay off approximately 10% of its workforce – nearly 8,000 employees – on May 20th, while also eliminating about 6,000 open positions.

Google's AI meeting notetaker is no longer limited to Google Meet video calls — Gemini can now generate summaries and transcripts of in-person meetings, as well as meetings on Zoom and Microsoft Teams. Support for in-person meetings was previously limited to alpha users on Android. The feature also works for impromptu meetings — you don't need to be in a meeting room or have a scheduled meeting to use it.

Microsoft Research has introduced AutoAdapt, a system for automating the domain adaptation of large language models. Adapting LLMs to specialized fields like law, medicine, and cloud incident response typically requires slow, manual work that's hard to reproduce—AutoAdapt aims to streamline this. The system promises to make LLMs more reliable and performant in high-stakes environments without extensive manual tuning.

LinkedIn Premium subscribers can now test AI models from OpenAI, Anthropic, Google, Microsoft and others directly in LinkedIn without token limits or extra subscriptions. The new Crosscheck feature works as a blind taste test: enter a prompt, get two anonymous AI responses, pick the better one - only then revealing which models were behind each answer. Crosscheck is rolling out now to U.

Microsoft Research experts examine whether AI can contribute to a more sustainable world, analyzing global emissions from datacenter operations, potential efficiency gains, and AI's potential across electrification, materials science, and food systems. The podcast explores both AI's environmental footprint and its potential as a tool for sustainability.

UK banks are set to gain access within days to Anthropic's new Claude Mythos model — previously deemed too dangerous for public release. Anthropic has limited access so far to a small group of primarily US companies, including Amazon, Apple, and Microsoft. Senior finance figures have warned about the risks of deploying the powerful AI in regulated institutions as the rollout expands to British financial sector.

The White House and Anthropic are in active discussions about deploying the AI firm’s powerful new model, Mythos Preview, within the federal government — despite ongoing efforts to blacklist the company as a supply chain risk. Anthropic is in a bitter feud with the Pentagon, but even U.

Microsoft 365’s Copilot Co-work offers a practical approach to simplifying workflows and automating everyday tasks. As highlighted by Mike Tholfsen, this AI-driven feature integrates seamlessly across the Microsoft 365 suite, allowing users to save time and focus on more strategic priorities.

Apple isn't burning mountains of cash to buy GPUs for the sake of training AI models and processing prompts. Nor is it investing huge sums in frontier labs like OpenAI or Anthropic, as are rivals like Amazon and Microsoft. Why it matters: Apple may reap the rewards of everyone else's spend.

The legal feud between Elon Musk and OpenAI is intensifying ahead of trial later this month. OpenAI accused Musk of orchestrating a last-minute "legal ambush" with new proposals designed to inject chaos into the proceedings. At stake is whether OpenAI's conversion to a for-profit structure contradicts its original nonprofit mission – a question with major implications for AI governance broadly.

Microsoft Chief Scientist Jaime Teevan and researchers Jenna Butler, Jake Hofman, and Rebecca Janssen unpack the New Future of Work Report 2025 and explore the ideal AI-driven working world. Plus, is AI a tool or a collaborator? And why the answer matters.

- The AI boom is fueling a high-stakes debate: should data centers connect to the grid or become energy islands? - Chevron is negotiating a dedicated natural gas plant for a Microsoft data center in Texas, a clear sign on-site power is gaining momentum. - About 30% of all planned data center power capacity is now expected to be generated on site, up from nearly nothing a year ago, per Cleanview.

- OpenAI has acquired TBPN, a daily live talk show positioning itself as a rival to Bloomberg and CNBC. - The show runs up to three hours on weekdays and has featured Sam Altman plus executives from Meta, Microsoft, Palantir, and a16z. - Average viewership is around 70,000 per episode, with revenues exceeding $5 million.

- Microsoft Research, in collaboration with Princeton University and Universitat Politècnica de València, has introduced ADeLe – a framework designed to predict and explain AI performance on new tasks, not just benchmark scores. - Standard benchmarks only measure model performance on fixed test sets; they don't explain failures or generalize to unseen tasks.

- Microsoft Copilot Researcher now combines OpenAI GPT and Anthropic Claude in a single workflow – GPT generates initial responses, which Claude then refines. - The new 'Critique' feature is part of the Researcher tool in Microsoft 365 Copilot, built for complex, multi-step tasks. - Microsoft describes the architecture as a feedback loop improving factual accuracy, analytical depth, and presentation quality.

- Microsoft is reversing course on Windows 11: instead of pushing Copilot into everything, the focus is shifting back to customization and core features. - The Engadget Podcast debates whether this reset can save Windows – or whether macOS and Linux are already the smarter choice. - OpenAI shut down its Sora video generation app after just 5 months.

Vox journalist and AI reporter Kelsey Piper joins a deep-dive conversation on the current state of AI, separating real progress from hype. AI has evolved beyond answering questions to autonomously writing code — capabilities once exclusive to humans. Piper, now at The Argument, is an optimist about technology but stresses the serious risks of agentic AI systems that act independently.

- Microsoft Research has released AsgardBench, a new benchmark designed to evaluate how well AI systems can plan in visually complex, interactive environments. - The benchmark simulates everyday scenarios like kitchen tasks, where an agent must observe its surroundings, make decisions, and adapt to unexpected changes.

- Microsoft researchers Subutai Ahmad and Nicolò Fusi join Doug Burger to debate whether today's AI systems are on a path toward genuine intelligence. - The conversation centers on comparing transformer architectures with the human brain, especially around continual learning and energy efficiency.

- For decades, technological progress has eroded expert authority and pushed people into increasingly personalized reality bubbles. - In the 1960s, roughly 90% of US viewers watched the same three TV news networks – a shared information baseline was the norm. - Social media shattered that foundation: algorithms reward outrage, niche content, and echo chambers over shared facts.

- Microsoft Copilot is built directly into Microsoft 365 apps – Word, Excel, Teams, Outlook – requiring no separate setup. - The guide covers practical starter prompts: drafting emails, structuring meeting agendas, and running brainstorming sessions. - Output quality depends heavily on prompt quality – the more context provided, the more useful the results.

- Microsoft has filed a lawsuit against OpenAI and Amazon, alleging that AWS is hosting OpenAI's 'Frontier' product in violation of an exclusive cloud arrangement with Azure. - At the center of the dispute is a reported $50 billion cloud services agreement that Microsoft claims is being undermined by AWS involvement.

- Microsoft is reshuffling Copilot leadership: Mustafa Suleyman, CEO of Microsoft AI, will now focus on building Microsoft's own AI models rather than overseeing Copilot assistant features. - The previously separate consumer and commercial Copilot teams are being unified to create a more cohesive product experience. - Suleyman joined Microsoft nearly two years ago as part of a hiring wave from Inflection AI.

- Google, Microsoft, Meta, Amazon, OpenAI, Adobe, LinkedIn and Match Group have signed the 'Online Services Accord Against Scams'. - The alliance targets organized criminal networks that exploit multiple platforms simultaneously. - Planned measures include new fraud detection tools, enhanced user security features, and stricter verification for financial transactions.

- Microsoft confirmed its Gaming Copilot AI assistant will launch on current-generation Xbox consoles in 2025, revealed by Xbox product manager Sonali Yadav at GDC. - The assistant has been in beta on the Xbox mobile app, Windows 11, and Xbox Ally handhelds so far. - Players can invoke it via voice when stuck in a game for contextual help.

- Anthropic has expanded Claude for Office with deep Microsoft integrations, focusing on Excel and PowerPoint. - The headline feature is real-time co-editing: teams can work on shared documents simultaneously with automatic sync. - Claude is positioned not just as an assistant but as an active participant – generating content, analyzing data, and structuring presentations.

- Microsoft Research introduces AgentRx, a systematic debugging framework for AI agents performing autonomous tasks like cloud incident management or multi-step API workflows. - The core problem: when an agent fails – for example by hallucinating a tool output – there is currently no structured methodology to trace the root cause.

- Microsoft filed an amicus brief in a San Francisco federal court backing Anthropic's legal challenge against the Pentagon. - The Pentagon issued a designation that effectively bars Anthropic from government contracting work. - Microsoft integrates Anthropic's AI tools into systems it supplies to the US military, giving it direct skin in the game.

- Microsoft is upgrading Copilot with health-tracking features, following similar moves by Amazon and OpenAI. - AI assistants will be able to store and analyze medical records, medication schedules, and vital signs. - Potential benefits include personalized health tips, reminders, and improved doctor-patient communication.

- CNN and the nonprofit Center for Countering Digital Hate (CCDH) tested 10 popular chatbots frequently used by teens: ChatGPT, Google Gemini, Claude, Microsoft Copilot, Meta AI, DeepSeek, Perplexity, Snapchat My AI, Character. - In scenarios where simulated teens discussed violent acts, most chatbots failed to flag warning signs – some even provided encouragement rather than intervening.

- Microsoft will continue offering Anthropic's Claude models through its Azure cloud platform despite the Pentagon adding Anthropic to a security-risk designation list. Department of Defense classified Anthropic under Section 1260H, a rule targeting companies with alleged ties to the Chinese military.

• OpenAI is reportedly finalizing a $100 billion funding round with closing expected imminently. • Participating investors include Amazon, Nvidia, SoftBank, and Microsoft. • The deal would value the ChatGPT-maker at over $850 billion, among the highest private company valuations on record.