Openhour
The weekly AI brief  ·  July 2026

Claude Goes to School

Anthropic builds a Claude for teachers while OpenAI locks down ChatGPT for teens

This week the big AI labs stopped talking to engineers and started talking to teachers, parents, and kids. Anthropic launched a version of Claude built for classrooms, OpenAI rolled out teen safety controls, and Google quietly turned its search into an assistant that gets things done. Here is what matters if you build or run anything.


ANTHROPIC

Claude shows up for teachers

Anthropic launched Claude for Teachers, a version of its AI made for the classroom. The idea is simple: help teachers plan lessons, build materials, and save time without doing the thinking for students.

This is Anthropic leaning into education as a serious use case, not a side project. It also fits their whole pitch of building AI that is careful and trustworthy, which matters a lot more when the users are kids.

Separately, Anthropic put $10 million into Canadian research institutions to fund the next wave of AI work. Small dollar figure by tech standards, but a signal about where they want the field to grow.

Why it matters If AI can help a busy teacher plan a week of lessons, it can help you draft, plan, and organize the repetitive parts of your own work.

OPENAI

ChatGPT gets guardrails for teens

OpenAI announced new protections aimed at younger users. That means age-appropriate settings, learning tools built for students, parental controls, and partnerships with outside experts.

The message is that safety is now a product feature, not just a policy page. As more families and schools use these tools, the companies that make trust easy will win the adoption battle.

OpenAI also laid out its view on how AI should be regulated, calling for state and federal rules to work together. Worth watching if your business touches anything regulated.

Why it matters Clear safety controls make AI easier to bring into schools, families, and any customer-facing product you run.

TOOLS

Google search stops answering and starts doing

Google's AI Mode can now connect to select apps and actually complete tasks, not just answer questions. Think of it moving from a smart librarian to a helpful assistant that presses the buttons for you.

This is the shift everyone keeps calling the "agentic era" in plain terms: AI that takes actions across the tools you already use. Booking, ordering, editing, all from one place.

For small teams, this is the direction to plan for. The winners will be the ones who let an AI do the boring steps between deciding and done.

Why it matters When AI can finish tasks across your apps, you spend less time clicking and more time on the work that actually needs you.

MARKETING

Star in your own AI videos

Google Vids now lets you create videos featuring a digital version of yourself, a personal AI avatar. Alongside it, Google added Gemini-powered tools to generate and edit videos from a text prompt or a reference image.

For anyone who avoids being on camera, this is a real shortcut. You could record once, then produce clips, updates, and short explainers without setting up lights again.

Quality still varies, so treat it as a first draft tool. But the barrier to making video just dropped hard.

Why it matters Video content that used to need a crew can now be made from a prompt, which levels the field for solo operators and small brands.

AGENTS

You can now order DoorDash from a terminal

DoorDash opened a beta of dd-cli, a command-line tool that lets developers and AI agents search stores, build carts, and place orders. In plain terms, it is software built so an AI can shop for you instead of a human tapping through an app.

This is a small headline with a big hint. Companies are starting to build products designed for AI agents to use, not just people.

Expect more of this. The businesses that make it easy for an agent to buy, book, or sign up will quietly get more sales.

Why it matters If AI agents become the customers, the way you design your checkout and signup flow may need to change.

WEB

AI bots now need permission to visit

Starting September 15, a slice of the web will block AI agent crawlers by default. These are the bots that fetch a page in real time when someone asks an AI a question. Cloudflare announced the change, and it puts the choice back in site owners' hands.

If you run a website, this affects whether AI tools can read and cite you. Being blocked means invisible to the AI answers people increasingly rely on.

The fix is to set your permissions on purpose rather than by accident. Decide who you want in.

Why it matters Your website's visibility in AI answers now depends on rules you may need to set yourself.


Try this week

Pick one repetitive task you do every week, like writing a recurring email or planning a schedule, and ask Claude or ChatGPT to build you a reusable template for it. Save the prompt so you never start from scratch again.

Which of these are you actually going to try? Hit reply and tell me. I read every response and often turn them into next week's issue.

Free · every Friday

Get the weekly AI brief

Plain-English AI, in your inbox each week. One email, no spam.

No spam, ever. Unsubscribe with one click.


Sources
  1. Introducing Claude for Teachers anthropic.com
  2. Why teens deserve access to safe AI openai.com
  3. Google's AI Mode now lets you link and interact with select apps techcrunch.com
  4. Google Vids now lets you star in your own AI videos techcrunch.com
  5. Yes, you can now order DoorDash from the command line techcrunch.com
  6. AI agent crawlers now need permission. Here's how to get it artificialintelligence-news.com