Gmail privacy concerns, GPT group chats & more (November 21, 2025)

Good morning,

AI is moving into your group chats, inbox, car dashboard, and even state policy battles. Today’s updates show AI becoming more collaborative, more embedded, and more controversial across privacy, regulation, and enterprise stacks.

Let's dig in👇

🤝 Collab, Cars & Creative AI

💬 ChatGPT group chats arrive for everyone

ChatGPT rolls out group chats globally for up to 20 users, making AI a shared space instead of a solo tool. This positions ChatGPT as a collaborative hub that competes with Slack, Notion, and Teams in real time creativity and coordination.

🖼️ Google’s Nano Banana Pro supercharges Gemini visuals

Google launches Nano Banana Pro image capabilities that speed up branded visuals for marketing teams inside the Gemini 3 update. The expansion makes Google’s AI creative suite more competitive with Adobe and Canva as enterprise workflows shift toward prompt based content creation.

🚗 Gemini hops into Android Auto

Google’s Gemini arrives on Android Auto, letting drivers handle directions, messaging, and planning through conversational voice. It further embeds Google’s AI across daily routines while raising new questions around in car attention and safety.

🛡️ Models, Privacy & Power Plays

🛡️ Gmail quietly opts users into AI training

Gmail is now using inbox emails and attachments to train its AI unless users manually opt out. This shift intensifies privacy concerns for anyone handling sensitive data and forces businesses to review their email practices.

🧩 Microsoft Foundry brings a frontier model buffet

Microsoft Foundry expands its agent service with access to Anthropic, Meta, DeepSeek, xAI, Phi, and more. It strengthens the “multi model ecosystem” trend where enterprises mix specialized models under one agent orchestration layer.

🏛️ Trump’s AI order targets state regulations

Trump’s new AI executive order threatens to pull federal broadband funds if states enforce rules judged too restrictive. It escalates the federal versus state fight over how AI outputs should be regulated, especially around discrimination requirements.

🧰 Tools of the Day

→ Lamatic AI: Build reliable process driven agents that follow your internal workflows.
→ OCR Arena: Compare OCR engines on the same docs to pick the most accurate.
→ Globe of History: Explore 6,000 years of global history in interactive 3D.

💡 Prompt of the Day

“You are my ‘AI traffic controller.’ Ask me five questions about my current work, then design a weekly workflow that summarizes updates, drafts key replies, and surfaces the five actions I should take each day.”

⚡ Quick Hits

→ OpenAI debuts GPT-5.1: Codex Max shows serious long duration coding power.
→ Wispr lands $25M: Its voice dictation app continues rapid growth.
→ Meta’s SAM 3D: A major shift toward real world perception.
→ Regulators eye AI: Lobbying pressure builds across industries.
→ Major labels back Klay: AI personalized streaming gets heavyweight support.
→ TCS taps TPG: Funding secured for India’s new $2B AI data center.

🧾 TLDR

Group chats make ChatGPT a shared workspace, Google expands Gemini from images to car dashboards, and Foundry doubles down on multi model orchestration. Gmail’s quiet opt in for AI training raises privacy alarms, and Trump’s executive order sets up a major state versus federal showdown over AI rules.

Closing,
David