← Back to archive
Microsoft drops OpenAI for homegrown Office AI

Image from a featured story

Wednesday, July 8, 2026

Microsoft drops OpenAI for homegrown Office AI

Microsoft is replacing OpenAI and Anthropic models with its own AI in Office apps. Meta launched Muse Image generation across its social platforms and is previewing Content Seal, an invisible watermarking system that can detect AI-generated images even after edits and screenshots.

Top Stories

1
Microsoft Replaces OpenAI and Anthropic Models

Bloomberg

Microsoft is transitioning Office apps away from OpenAI and Anthropic models to proprietary in-house AI, marking a strategic pivot toward self-sufficiency in enterprise AI capabilities.

microsoftopenaianthropicin-house-models
2
Anthropic Expands Claude Cowork to Web and Mobile

claude.com

Anthropic's Claude Cowork is now available on web and mobile with background task execution and cross-device continuity, addressing the finding that 90% of agent usage is for business operations and content creation rather than coding.

anthropicagentsclaudeenterprise-ai-adoption
3
Introducing Muse Image: Image Generation Built for Your World

Meta

Meta has released Muse Image, its first image generation model integrated into Meta AI, enabling users across Instagram, WhatsApp, and other Meta apps to create and edit high-quality images using conversational prompts, with features like text rendering, Instagram account tagging, and upcoming advertising integration.

metaimage-generationgenerative-aimultimodal
4
Claude Fable 5 promotional access

Anthropic Support

Anthropic is providing limited-time promotional access to Claude Fable 5 for paid subscribers from July 1-12, 2026, allowing them to use up to 50% of weekly limits at no extra cost before switching to usage credits or alternative models.

anthropicclaudellmproduct-launch
5
Meta is previewing an invisible watermarking system called Content Seal

Engadget

Meta launched Content Seal, an invisible watermarking and detection system for AI-generated images that survives common modifications, marking a new approach to content authentication amid industry scrutiny over AI disclosure practices.

watermarkingcontent-authenticationai-detectiongenerative-ai

Keep Reading

Enjoyed this issue?

Get daily AI intel delivered to your inbox. No fluff, just the stories that matter.