Anthropic launched Claude Academy today. It's free, open to everyone, and accessible without logging in. The URL is academy.claude.com.

Courses are divided into two sections. The first covers Claude products: Claude.ai, Claude Cowork, Claude Code, Claude Tag, and Claude Platform. The second is a foundational AI literacy program called AI Fluency, which covers what AI can and can't do, and how to collaborate effectively with AI. It's worth checking out even if you don't use Claude.


The platform includes structured courses, tutorials, use cases, plus quizzes and completion badges. It offers personalized recommendations based on your interests and lets you track learning progress. If you want Claude to recommend courses tailored to your work style, you can install the official Skill.

The truly interesting part is Anthropic's teaching philosophy. Instead of teaching you more prompt engineering tricks, it focuses on durable mental models. One example is the principle: "The AI you use today is the worst AI you'll ever use in the future" — which means technology advances rapidly, so there's no point in memorizing static use cases. Another is "verify according to risk level" — let AI handle low-risk tasks freely, but review high-risk outputs yourself.


They even specifically note that where you once had to tell Claude "my audience is my legal colleagues", modern models will proactively ask for that context. So teaching specific workflows is useless — teaching good judgment is what matters.

This is where it defies conventional wisdom: The goal of this AI education isn't to offload everything to AI. Anthropic emphasizes "augmenting human agency", and courses guide learners to think through which tasks should go to AI, and which you should do yourself. For example, you might draft sensitive memos yourself, then have AI make a summary slide; let AI handle exploratory data analysis, but do the final check yourself. Also, you should always disclose AI use clearly when you rely on it, letting others know which parts of your work were generated with AI.

This approach is pretty rare right now. Most AI education teaches how to use AI to get more done, but Anthropic is teaching "when not to use it". They even remind learners not to let their own skills atrophy — some abilities are worth keeping sharp.

Claude Academy's course design also emphasizes hands-on practice. Every tutorial encourages you to work with Claude as you learn. The official blog notes that "today's Academy is the most rigid version it will ever be", because they'll use Claude itself to build personalized learning experiences in future updates.


How to access it: Just go to academy.claude.com directly, or click "Learn more" in the Claude profile menu. All courses are completely free.

Commenters have already asked about certificates, when German language support will be added, and whether access grants free Claude credits. The official team hasn't responded yet. But at the very least, an AI company putting serious effort into education is a good sign in itself.