Google Cloud is rolling out Gemini, a new AI platform built specifically for healthcare. The tool connects electronic health records with CMS datasets, giving providers a unified view of patient information. It also handles complex multi-step tasks like records synthesis and prior authorization, areas where clinicians reportedly lose about 30 minutes per patient interaction. Launch partners include HCA Healthcare, Highmark Health, and Northwell Health. The platform supports HIPAA compliance and is designed to automate routine work so staff can focus more on patient care.
Adobe is expanding its AI agent tools across Creative Cloud, adding MCP integrations that let AI agents assist with creative workflows. New features include a plan mode in Firefly and always-on AI agents for image quality checks. Adobe is also collaborating with Microsoft and Disney on new integrations, while emphasizing that safety and security remain top priorities. Meanwhile, the Frontier Model Forum released a framework to grade AI jailbreaks, rating them by capability gain, scope, ease of deployment, and ease of discovery. Anthropic, Google, Microsoft, and Amazon supported the development of this system, and Version 1.0 is open for public comment until December 21.
In other developments, CrowdStrike is acquiring Fiddler AI, a company that monitors how AI systems behave. Fiddler checks AI outputs for problems like hallucinations and private data exposure. The deal signals that cybersecurity is expanding to cover agentic AI that makes decisions and accesses enterprise information. At the same time, a new analysis argues that success in defense AI depends less on the model itself and more on the harness, meaning the software infrastructure, guardrails, and data systems surrounding it. Without proper data formatting and orchestration, spending on AI models yields limited returns.
Sierra co-founder Bret Taylor says the US must maintain leadership in AI despite public pushback against data centers. He argues that leadership in power, compute, chips, models, and applied AI is essential. Microsoft CEO Satya Nadella's vision of a future with millions of AI agents resurfaced after the company offered voluntary exit plans to about 60,000 employees. Nadella had previously imagined 20,000 humans working alongside millions of AI agents. Carnegie Mellon University is also investing heavily in AI research through partnerships with Google and Amazon, while building a new data center with the University of Pittsburgh.
Key Takeaways
- Google Cloud launched Gemini for healthcare, an agentic AI platform supporting HIPAA compliance and designed to automate tasks like prior authorization and records synthesis.
- Clinicians reportedly lose about 30 minutes per patient interaction on manual work that Gemini aims to reclaim for patient care.
- HCA Healthcare, Highmark Health, and Northwell Health are launch partners for Google's Gemini healthcare tool.
- Adobe is adding MCP integrations across Creative Cloud, including a Firefly plan mode and always-on AI agents for image quality checks, with Microsoft and Disney collaborations in progress.
- The Frontier Model Forum, supported by Anthropic, Google, Microsoft, and Amazon, released a jailbreak rating framework open for public comment until December 21.
- CrowdStrike is acquiring Fiddler AI to monitor AI system behavior, signaling a shift toward cybersecurity covering agentic AI that makes enterprise decisions.
- A new analysis argues defense AI success depends more on the harness, including data infrastructure and guardrails, than on the AI model itself.
- Sierra co-founder Bret Taylor says the US must lead in AI across power, compute, chips, models, and applied AI despite public concerns about data centers.
- Microsoft offered voluntary exit plans to about 60,000 employees, coinciding with a resurfaced video of Satya Nadella envisioning millions of AI agents working alongside humans.
- Carnegie Mellon University is partnering with Google and Amazon on AI research and building a new data center with the University of Pittsburgh.
Google Cloud launches Gemini AI for healthcare data
Google Cloud launched Gemini, an AI tool built for healthcare. It connects data from electronic health records and CMS datasets. The tool gives providers a unified view of patient information. Google Cloud is partnering with healthcare groups to test and refine Gemini. It aims to improve patient care and outcomes.
Google Cloud launches Gemini for healthcare workflow help
Google Cloud introduced Gemini for healthcare to handle complex multi-step tasks. Clinicians lose about 30 minutes per patient interaction on manual work like records synthesis and prior authorization. Gemini works within guardrails and includes human oversight. Launch partners include HCA Healthcare, Highmark Health, and Northwell Health. The tool is designed to reclaim staff time for patient care.
Google Cloud rolls out Gemini healthcare AI platform
Google Cloud rolled out Gemini, an agentic AI platform for healthcare. It was developed with clinical and health plan leaders and supports HIPAA compliance. The platform helps organizations automate tasks, improve patient outcomes, and reduce costs. Dr. Brian Brugger called it a game-changer for healthcare organizations. It uses machine learning to analyze data and support better decisions.
Defense AI success depends on harness more than model
A new analysis argues that super intelligence success depends more on the harness than the AI model. The harness includes software infrastructure, tools, guardrails, and data systems that surround the AI engine. Defense and intelligence agencies face challenges because AI tools can fail or give unreliable results. The author says data must be properly formatted and machine-readable. Without good data structure and orchestration, spending on AI models yields limited returns.
Sierra co-founder says US must lead in AI build-out
Sierra co-founder Bret Taylor says the US must lead in AI despite public pushback. He spoke about data centers being on the ballot after midterm elections. Protesters have raised concerns about noise, electricity costs, and environmental impact. Taylor argues AI is too important and the country will navigate the risks. He says leadership in power, compute, chips, models, and applied AI is essential.
Nadella's AI future vision resurfaces as Microsoft offers 60,000 workers exit
A video of Microsoft CEO Satya Nadella from June 2026 resurfaced on social media. In it, he imagined 20,000 humans working with millions of AI agents. This comes after Microsoft offered a voluntary exit plan to about 60,000 employees. The offer includes financial support and career help for eligible workers in the US, UK, Canada, and India.
CMU and Pitt adopt AI as data center debate grows
Carnegie Mellon University is investing heavily in AI research and partnerships with Google and Amazon. University President Farnam Jahanian shared these plans at a Washington summit. The school is also building a new data center with the University of Pittsburgh. Leaders say universities must balance AI needs with community and environmental concerns.
Adobe expands AI agent tools across Creative Cloud
Adobe is adding more MCP integrations so AI agents can help with creative tasks. New features include a plan mode in Firefly and always-on AI agents for image quality checks. The company is also working with Microsoft and Disney on new integrations. Adobe says safety and security remain top priorities in its AI partnerships.
AI takes center stage at American Film Market
The 2026 American Film Market in Los Angeles is putting AI front and center. The Innovation Hub is expanding to showcase AI tools used in filmmaking. Companies like Largo.ai and Oxen.ai are showing how AI can help analyze film projects and create AI shots. Industry leaders say AI is mainly helping with data analysis and efficiency.
Frontier Model Forum proposes CVSS-style rating for AI jailbreaks
The Frontier Model Forum released a new framework to grade AI jailbreaks. It rates jailbreaks by capability gain, scope, ease of deployment, and ease of discovery. The system was created with support from Anthropic, Google, Microsoft, and Amazon. Version 1.0 is still open for public comment until December 21.
AI reshapes MBA programs at top Indian business schools
Business schools including IIM Kozhikode, IIM Lucknow, ISB and BITSoM are changing their MBA programs because of AI. Students must now build AI agents and learn how to work with machines across subjects like operations, marketing and decision-making. Schools are also adding AI labs, new courses in data analytics and stronger focus on ethics and responsible use of AI. The goal is to prepare graduates to lead in workplaces where AI handles routine tasks.
CrowdStrike Fiddler deal points to stronger AI security
CrowdStrike is moving to acquire Fiddler AI, a company that watches how AI systems behave. Fiddler, started in 2018, checks AI outputs for problems like hallucations and exposure of private data. The deal shows that cybersecurity is expanding to cover agentic AI that makes decisions and accesses enterprise information. Experts say AI security now needs constant monitoring, human oversight and automated controls to prevent failures.
Sources
- Google Cloud launches Gemini AI tool for healthcare
- Google Cloud launches Gemini for healthcare.
- Google Cloud rolls out Gemini for healthcare
- Why the harness matters far more than the model for super intelligence success
- AI is not popular right now, but the US has to lead: Sierra co-founder Bret Taylor
- Satya Nadella’s old AI workforce vision video resurfaces after 60,000 Microsoft employees offered exit scheme — WATCH
- CMU, Pitt embrace AI as region debates data centers
- Adobe Is Giving Its AI Agents Even More to Do
- AI Takes Center Stage at Tech-Forward AFM With Innovation Hub
- New framework grades AI jailbreaks across capability gains, scope, and how easy they are to deploy — Frontier Model Forum proposes CVSS-style rating system for new AI models
- AI pushes B-schools to rethink curriculum, skills and leadership training
- CrowdStrike–Fiddler Deal Signals New Era of AI Security
Comments
Please log in to post a comment.