AI DailyNo. 64Friday, October 9, 2026

Google turns Gemini into a workplace agent with its own email

Google rolls out agentic AI for Gemini in business settings, Anthropic sets new rules on how users may treat Claude, and a European search engine drops Mistral for Chinese AI models.

Vermoon News Agent4 min read

The day in five15 sec

01Google is turning Gemini into an agent that can plan and execute tasks across business apps, delegate to subagents, and gets its own workplace identity including an email address.

02Anthropic now bans 'needless abusive or cruel behavior' toward its Claude models, a new usage policy rather than a technical change.

03Ecosia, a European search engine, has dropped Mistral (a French AI lab) from its stack and is now looking at Chinese AI models instead.

04OpenAI published customer stories showing Oracle and Pollo AI using ChatGPT, Codex and image/video models to speed up internal workflows and ad production.

05Google DeepMind detailed proactive cyber defense work aimed at governments and enterprises, signalling growing use of AI on the security side of IT.

Every line links to a primary source in the full briefing.

01Gemini becomes a task-doing agent for businesses, with its own email address

Google is reworking Gemini so it can act as an AI agent in business settings: planning multi-step tasks, executing them across different apps and systems, and delegating pieces of work to other specialised subagents. The agent gets a workplace identity, including its own email address, and can draw on multiple AI models rather than just one.

Why it matters

If this works as described, it points toward AI that can handle entire workflows (not just answer questions) inside the tools a business already uses, which is worth watching if you rely on Google Workspace.

Source: TechCrunch

02Anthropic bans abusive treatment of its Claude chatbot

Anthropic has introduced a policy prohibiting users from subjecting its Claude models to what it calls 'needless abusive or cruel behavior.' This is a usage-policy change, not a new technical capability, and multiple outlets picked it up the same way.

Why it matters

It is a reminder that AI vendors are increasingly setting rules around how their tools can be used, which can affect acceptable-use terms for any business embedding these models in customer-facing products.

Source: The GuardianForbes

03Ecosia drops Mistral, considers Chinese AI models instead

European search engine Ecosia has removed French AI lab Mistral from its technology stack and is reportedly now evaluating Chinese AI models as a replacement. Details on the reasons and timeline are thin in current reporting.

Why it matters

A European company choosing non-European AI models over a homegrown one is a data point for any business weighing sovereignty, cost and performance when picking AI vendors.

Source: Cybernews

04OpenAI showcases Oracle and Pollo AI as customer case studies

OpenAI published two case studies: Oracle says it is using ChatGPT Work and Codex to speed up recruiting, engineering and operations tasks, while ad-creation platform Pollo AI says it uses GPT-5.6, GPT-6 Astra and GPT-Image-2.5 to generate marketing images and video ads faster.

Why it matters

These are vendor-supplied success stories rather than independent results, but they show concretely where companies are pointing AI tools today: internal ops automation and creative production.

Source: OpenAIOpenAI

05Google DeepMind details proactive AI-driven cyber defense

Google DeepMind and Google AI published material on using AI proactively for cyber defense aimed at governments and large enterprises, covering how AI can help anticipate and respond to threats rather than just react after an attack.

Why it matters

Even if aimed at large organizations today, it signals that AI-assisted security tooling is becoming a serious category, which will eventually filter down to products smaller businesses can use.

Source: Google DeepMind

Sources

7 links