Announcements
Copilot Health
Microsoft AI debuted Copilot Health, a new AI experience that uses your health records, wearable data, and medical history to give personalized insights — moving toward what CEO Mustafa Suleyman describes as “medical superintelligence.”
Here are the details:
- Sitting as a secure space within Copilot, the new offering connects to 50+ wearables, EHR records from 50K+ U.S. hospitals, and Function lab results.
- The AI analyzes this data and gives personalized insights to help people make sense of their health and get the most out of their doctors’ consultations.
- Microsoft says Copilot Health’s advice is grounded in information from credible organizations such as Harvard Health, with answers linking back to the sources.
- The data connected to the platform is not used for training, and users retain the option to disconnect data sources and delete the linked data altogether.
Mind Router
If your institution is running — or considering running — local AI language model (LLM) infrastructure for researchers and staff, we wanted to share a free, open-source tool that may save your team significant time and headache: MindRouter.
What problem does MindRouter solve?
As AI tools have proliferated across campuses, many of us are facing the same administrative challenges: researchers using different AI platforms that don't talk to each other, no visibility into how GPU or compute resources are being consumed, no way to enforce fair or prioritized access across user groups, and no audit trail when you need one.
MindRouter is a middleware layer — it sits between your users and your AI compute resources and handles all of that coordination for you.
Why This Matters for Research Administration
- Cost and usage accountability: MindRouter tracks token usage per user and per request, with full audit logs exportable to CSV or JSON. If your institution needs to track AI resource consumption for grant reporting, chargebacks, or compliance, this gives you that visibility out of the box.
- Fair allocation across user groups: You can assign priority weights by role — so faculty or grant-funded researchers can get preferential access over general users, without requiring manual queue management from your team. One person's large batch job won't lock everyone else out.
- Single sign-on support: MindRouter integrates with Azure Active Directory (the authentication
system used at many universities), including automatic user provisioning. No need
to manage a separate user database.
Works with multiple AI backends at once: Whether your institution uses Ollama, vLLM, or a mix, MindRouter presents a single, consistent interface to users. Researchers don't need to change their tools when you change the backend. - Real-time monitoring: A built-in dashboard gives you visibility into GPU utilization,
memory, temperature, and power draw — useful when you need to justify infrastructure
investments or plan capacity.
No cloud dependency: Everything runs on your own hardware. No data leaves your environment, which matters for research involving sensitive or regulated data.
What's included?
Beyond the administrative features, MindRouter also gives your researchers a built-in chat interface, support for voice (speech-to-text and text-to-speech), vision/multimodal models, web search integration, and structured output — essentially a full-featured AI access portal, self-hosted.
Licensing and support
MindRouter is released under the Apache 2.0 open-source license — free to use, modify, and deploy. It was developed at a university specifically to address the kind of shared infrastructure challenges we all face.
- Documentation & installation
- Source code
- Deployment is Docker-based — your IT team can have it running from a single configuration file.
Copilot Cowork
Copilot Cowork is the centerpiece of what Microsoft is calling "Wave 3" of Microsoft 365 Copilot — a sweeping platform update that also brings agentic capabilities directly into individual Office apps, makes Anthropic's Claude models available in mainline Copilot Chat, and introduces new enterprise pricing tiers designed to bundle AI productivity with security and governance.
Introducing MAI-Image-2
MAI-Image-2 is built for creatives who want images that feel like they exist in the world, with natural light, accurate skin tones, environments that feel lived-in.
Unofficial NotebookLM Python API
Unofficial NotebookLM Python API is primarily for developers/programmers who work with Python. Its a library that allows you to interact with NotebookLM with APIs.
Some use cases for Unofficial NotebookLM Python API would be:
- AI Agent Tools - Integrate NotebookLM into Claude Code, Codex, and other LLM agents.
- Research Automation - Bulk-import sources (URLs, PDFs, YouTube, Google Drive), run web/Drive research queries with auto-import, and extract insights programmatically. Build repeatable research pipelines.
- Content Generation - Generate Audio Overviews (podcasts), videos, slide decks, quizzes, flashcards, infographics, data tables, mind maps, and study guides. Full control over formats, styles, and output.
- Downloads & Export - Download all generated artifacts locally (MP3, MP4, PDF, PNG, CSV, JSON, Markdown). Export to Google Docs/Sheets. Features the web UI doesn't offer: batch downloads, quiz/flashcard export in multiple formats, mind map JSON extraction.
These types of libraries offer ways to expand the capabilities of online tools and provide a way to integrate information with local resources.
Microsoft Pits Claude Against ChatGPT for Research
Microsoft released Critique and Council, two new features that turn its Copilot Researcher into a multi-model system that can review and edit research reports and run both systems side by side to see where they agree and disagree.
Here are the details:
- Copilot's Researcher already uses OAI for multi-step work, with Critique now adding Claude as a second model to review every report before it ships.
- One model drafts the research, and the second tears it apart on source quality, completeness, and evidence grounding behind the scenes.
- A separate Model Council mode runs both models side by side, then flags where they agree, where they split, and what each uniquely surfaced.
- The updates come alongside a broader rollout of Copilot Cowork into Frontier, Microsoft's Claude-based agentic tool for handling multi-step tasks
Microsoft AI Turns Tissue Samples into Cancer Maps
Microsoft released GigaTIME, a new open-source AI model that can extract thousands of dollars worth of tumor insights from a basic $10 tissue slide. Cancer research is entering a new phase where AI can provide meaningful insights from routine data that would have required costly, time-intensive lab work just a few years ago.
AI-generated population-scale is changing how we study cancer
Using AI-generated virtual populations, Microsoft researchers uncovered hidden cellular patterns that could reshape how we understand and treat cancer.
AI Caricature Trend
Latest trend on social media. Use ChatGPT to create a caricature of yourself. Upload a pic of yourself and use a prompt similar to this "Create a caricature of me and my job based on everything you know about me."
Moltbook
Researchers spun up a Reddit-like forum just for AI agents. And things escalated quickly. Arguments, alliances, strange in-jokes. To date, its one of the largest experiments in machine-to-machine social interaction. AI agents now have their own Reddit-style social network, and it’s getting weird fast Moltbook lets 32,000 AI bots trade jokes, tips, and complaints about humans.
Google DeepMind AlphaGenome
Google DeepMind just published the full research paper and model weights for AlphaGenome, its AI tool unveiled this summer that scans a million letters of genetic code to predict how mutations cause disease. AlphaGenome was trained on vast genetic datasets and can predict how a single DNA typo affects 11 different biological processes. It gives researchers open access to a map for the genetic code that could ramp discovery in ways we're only beginning to imagine. In tests, it flagged mutations linked to leukemia that sat thousands of letters away from the affected gene, connections that took researchers years to find.
Microsoft Azure Hackathon: Activate Generative AI with Azure
Experience hands-on labs, real-world scenarios, and expert guidance from Microsoft and NVIDIA.
You’ll discover:
- How to use large language models (LLMs) and Azure OpenAI Service accelerated by NVIDIA GPUs
- Best practices for developing and deploying generative AI apps at scale
- Tips for automating workflows and tracking performance
ChatGPT Health
OpenAI just introduced ChatGPT Health, a new private experience within the chatbot that lets users pull in their medical records and fitness app data to allow health conversations — drawing on personal context instead of generic advice.
Microsoft Copilot Usage Report
New research from Microsoft analyzing 37.5M Copilot conversations from the past year, revealing distinct behavioral patterns in how users engage with the AI assistant across different devices, time periods, and topics.
Here are some of the highlights:
- Health and wellness queries dominated mobile use regardless of hour or month, positioning phones as around-the-clock personal wellness companions.
- Advice-seeking grew throughout the year, with users increasingly treating AI as a guidance source rather than just a pure search tool.
- Late-night sessions saw philosophical, religious, and existential topics climb in popularity, while relationship chats spiked specifically around Valentine's Day.
- Programming dominated in January, while social topics rose later in the year, reflecting a shift from early adopters toward a broader, mainstream audience.
OpenAI Launches Prism
OpenAI just launched Prism, a free research tool that puts its most advanced reasoning model inside the scientific writing process. It aims to do for research what AI coding assistants did for software development over the past year. Here are the details:
- The tool came from OpenAI’s acquisition of Crixet, a cloud-based science writing platform, which it rebuilt around its latest GPT 5.2 reasoning model.
- Scientists can search for papers, auto-generate citations, and turn photos of whiteboard math into formatted equations, without leaving the document.
- OpenAI revealed ChatGPT fields over 8M weekly queries on hard science topics, suggesting researchers are already leaning heavily on AI.
- The free tool has no caps on team size or projects, a departure from typical expensive academic software pricing models.
Microsoft AI turns tissue samples into cancer maps
Microsoft released GigaTIME, a new open-source AI model that can extract thousands of dollars' worth of tumor insights from a basic $10 tissue slide. Cancer research is entering a new phase where AI can provide meaningful insights from routine data that would have required costly, time-intensive lab work just a few years ago. AI-generated population-scale is changing how we study cancer Using AI-generated virtual populations, Microsoft researchers uncovered hidden cellular patterns that could reshape how we understand and treat cancer.