microsoft, openai and anthropic Updates

Microsoft is making significant strides in AI, launching in-house models like MAI-Voice-1 for Copilot's audio features and MAI-1-preview for general text tasks. The MAI-1-preview is currently undergoing public testing on LMArena. This move signals a push for AI independence, even with Microsoft's $13 billion investment in OpenAI. Meanwhile, the General Services Administration (GSA) faces a protest over its $1 AI tool contracts with OpenAI, Anthropic, and Google, with Ask Sage CEO Nic Chaillan citing fairness and cybersecurity concerns. Meta anticipates modest sales of 150,000-200,000 units for its $800 Hypernova AI smart glasses over two years, viewing them as an experimental step toward AI-integrated wearables. In education, schools in Iowa and Massachusetts are integrating AI into lesson plans, aiming to enhance learning and critical thinking. Spokane Falls Community College will launch an AI certificate program in fall 2026, supported by a $152,000 grant. HealthEquity's AI tool, HSAnswers, is winning awards for its ability to simplify healthcare benefits education. Gartner predicts that AI will be integrated into one-third of enterprise apps by 2030, and 70% of finance functions will use AI analysis for real-time decisions by 2028, while also warning about potential isolation among finance workers. Jatin Sandilya is using AI to improve software integrations, with his AI SDK enabling AI systems to work with platforms like Salesforce and HubSpot. The Open Source Security Foundation (OpenSSF) is also addressing AI security, releasing a whitepaper and recognizing contributions at Community Day Europe. Overall, AI's increasing presence raises questions about trust and its impact on various sectors.

Key Takeaways

• Microsoft is rolling out its own AI models, MAI-Voice-1 and MAI-1-preview, for Copilot, signaling a move towards AI independence from OpenAI.
• Microsoft's MAI-1-preview is undergoing public testing on LMArena to gather user feedback on its text generation capabilities.
• The GSA's $1 AI tool contracts with OpenAI, Anthropic, and Google are facing protests over fairness and cybersecurity concerns.
• Meta projects modest sales of 150,000-200,000 units for its $800 Hypernova AI smart glasses, viewing them as an AI-integrated wearable experiment.
• Schools in Iowa and Massachusetts are integrating AI into lesson plans to enhance student learning and critical thinking skills.
• Spokane Falls Community College will offer an AI certificate program starting in fall 2026, funded by a $152,000 grant.
• HealthEquity's AI-powered tool, HSAnswers, is recognized for its innovation in simplifying healthcare benefits education.
• Gartner forecasts that AI will be integrated into one-third of enterprise apps by 2030, with 70% of finance functions using AI analysis by 2028.
• Jatin Sandilya is using AI to improve software integrations, enabling AI systems to work with platforms like Salesforce and HubSpot.
• The Open Source Security Foundation (OpenSSF) is actively addressing AI security through whitepapers and community recognition.

Microsoft unveils MAI-Voice-1 and MAI-1-preview AI models

Microsoft introduced two new AI models, MAI-Voice-1 for speech generation and MAI-1-preview for text. MAI-Voice-1 powers Copilot's Daily and Podcast features, while MAI-1-preview is available for public testing on LMArena. Microsoft focused on efficiency, training MAI-1-preview using fewer GPUs than other models like xAI's Grok. These models signal Microsoft's aim to compete independently in AI, despite its investment in OpenAI's GPT tech.

Microsoft tests MAI-1-preview its first in-house AI model

Microsoft's MAI division started public testing of MAI-1-preview, its first foundation model trained in-house. The model is being tested on LMArena and is designed for everyday questions. It will soon be used in Copilot for some text tasks. Microsoft also released MAI-Voice-1, a speech generation model for Copilot Daily and Podcasts. This move signals Microsoft's effort to reduce reliance on OpenAI, in which it has invested heavily.

Microsoft shifts from OpenAI with in-house AI for Copilot

Microsoft is moving towards AI independence by launching its own AI models, MAI-Voice-1 and MAI-1-preview, for Copilot. MAI-Voice-1 generates audio quickly and is used in Microsoft's news summaries and Copilot Podcasts. MAI-1-preview is being tested on LMArena for user feedback. Microsoft invested $13 billion in OpenAI, but hired Mustafa Suleyman and his team to develop AI models independently. Microsoft aims to have options and will continue using OpenAI models, but sees value in its own AI development.

Microsoft's new AI models compete with OpenAI's options

Microsoft revealed two AI models, MAI-Voice 1 for natural speech in Copilot and MAI-1-preview for general tasks, rivaling OpenAI. MAI-Voice 1 creates realistic voices and is used in Copilot Daily and Podcasts. MAI-1-preview, a general model, is undergoing public testing. Microsoft plans to gather feedback on MAI-1-preview's text generation within Copilot. These models signal Microsoft's competition with OpenAI, despite their partnership.

Iowa schools use AI in lesson plans for students

Schools in Iowa, like Paton-Churdan and Ankeny, are adding AI to their lesson plans. Teachers are using AI for tasks like planning lessons and giving feedback. They want students to use AI as a tool to boost creativity and learning, but not to replace thinking. Students have mixed feelings, with some excited about AI's potential and others worried about it being an easy way out of work.

Massachusetts schools explore AI benefits in classrooms

Schools in Massachusetts are exploring how to use AI in classrooms. The state's Department of Elementary and Secondary Education is providing guidance and training for teachers. Schools are finding ways to use AI as a tool to help students learn and think critically. They aim to avoid AI doing the work for students, like writing essays, and instead focus on using it to enhance learning.

OpenSSF honors AI security efforts at Community Day Europe

The Open Source Security Foundation (OpenSSF) celebrated achievements in AI/ML security and community contributions at Community Day Europe. They presented Golden Egg Awards to Ben Cotton, Kairo de Araujo, Katherine Druckman, Eddie Knight, and Georg Kunz for their work in open source security. OpenSSF released a whitepaper on securing the AI lifecycle and participated in the AI Village at DEF CON. They also co-launched initiatives for cybersecurity skills and global cyber policy.

GSA's $1 AI tool awards face protest over fairness

Ask Sage CEO Nic Chaillan is protesting the General Services Administration's (GSA) $1 contracts with OpenAI, Anthropic, and Google. He claims the awards didn't follow rules like the Competition in Contracting Act. Ask Sage argues the low prices may lock agencies into using those AI providers, limiting competition. Chaillan also raises concerns about cybersecurity and a lack of transparency in the contract terms.

Meta expects modest sales for Hypernova AI smart glasses

Meta forecasts selling 150,000-200,000 units of its $800 Hypernova AI smart glasses over two years. These glasses feature AI and a neural wristband. The low sales expectations reflect challenges in the augmented reality market. Meta views Hypernova as an experiment, prioritizing innovation and user feedback. The glasses may signal Meta's shift towards AI-integrated wearable devices.

Gartner warns AI could isolate finance workers

Gartner predicts AI will change finance, with one-third of enterprise apps using AI agents by 2030. They say 70% of finance functions will use AI analysis for real-time decisions by 2028. Gartner warns that AI and remote work could make finance staff more isolated, creating a 'lonely enterprise'. They also predict a shortage of finance talent as many CPAs retire. Gartner advises CFOs to manage these changes to support employees.

HealthEquity's AI tool wins awards for benefits education

HealthEquity's AI-powered tool, HSAnswers, won awards for its innovation in benefits education. HSAnswers uses AI to quickly answer questions about Health Savings Accounts (HSAs) and other benefits. It provides personalized information using HealthEquity's knowledge and expertise. Since launching in October 2024, HSAnswers has facilitated many conversations, making healthcare benefits easier to understand.

Jatin Sandilya uses AI for reliable software integrations

Jatin Sandilya helps companies connect different software programs using AI. He founded Revert, a company that made it easier for early-stage companies to test and ship integrations. Revert was acquired by Ampersand, where Sandilya became Head of AI. He built an AI SDK that lets AI systems work with platforms like Salesforce and HubSpot. Sandilya believes software is moving towards an 'agentic' model, where AI handles tasks automatically.

Spokane Falls CC offers AI certificate program

Spokane Falls Community College will offer a certificate in artificial intelligence starting in fall 2026. The program will teach students to use AI in cybersecurity, healthcare, and manufacturing. The college will also offer two-day classes for the public on using AI in work and life. The AI program is funded by a $152,000 grant from Washington State.

Do you trust AI to improve our lives?

AI is becoming more common in our lives, from language models to problem-solving algorithms. It's used in healthcare, education, and to fight climate change. But there's a debate about whether we can trust AI. This article asks readers to share their opinion on whether they trust AI to act safely and fairly.

Sources

• Microsoft introduces a pair of in-house AI models
• Microsoft Launches Public Testing of First In-House Foundation Model
• Microsoft’s signals shift from OpenAI with launch of first in-house AI models for Copilot
• Microsoft’s Two New AI Models Rival OpenAI's Similar Options
• Iowa schools incorporate AI into lesson plans
• How can AI be used in the classroom? Massachusetts schools are working to identify the benefits
• OpenSSF Celebrates Global Momentum, AI/ML Security Initiatives and Golden Egg Award Winners at Community Day Europe
• GSA’s $1 awards for AI tools come under protest
• Meta Forecasts 150K-200K Sales for $800 Hypernova AI Smart Glasses
• Gartner predicts an AI-fueled ‘lonely enterprise’ for finance workers if CFOs don't take action
• HealthEquity's AI-Powered Tool Recognized for Innovation in Streamlining Benefits Education
• Meet Jatin Sandilya: The Founder Driving Reliable Integrations With AI
• Spokane Falls CC's New AI Certificate Focuses On Cybersecurity, Health
• Do you trust AI?