I’m going to tell you something that might get me in trouble with my friends in Silicon Valley. Most of the advice out there about building an AI curriculum is garbage. It’s written by developers who think teaching is just about listing the hottest new tools. They’ve never stood in front of a classroom, never seen the blank stares when the jargon gets too thick, and never had to actually make this stuff work for real students.
I have. And I’m here to tell you that designing a great AI curriculum has almost nothing to do with learning how to use the latest AI model. It’s about something much deeper: changing the way we think, and the way we teach.
For years, I’ve been on both sides of the table. As an entrepreneur, I’ve built and sold companies like RemoteTeam and MovieLaLa. As an investor, I’ve been lucky enough to back over 200 startups, including some of the biggest names in AI like Anthropic, OpenAI, and Scale AI. But I’ve also spent countless hours in classrooms, talking to educators and students, trying to bridge the gap between the tech world and the real world. And what I’ve learned is that our approach to AI education is fundamentally broken.
We’re so obsessed with the “what”—the tools, the platforms, the programming languages—that we’ve completely forgotten the “why.” Why are we teaching this? What skills are we actually trying to build? The goal isn’t to turn every student into a machine learning engineer. The goal is to create a generation of critical thinkers who can use AI as a powerful tool for problem-solving, no matter what field they go into.
This is my counterintuitive guide to building an AI curriculum that does just that. It’s not about chasing trends. It’s about building a foundation of timeless skills that will serve students long after the current crop of AI tools has been replaced by the next big thing.
Principle 1: Stop Chasing the Tools
Every week, there’s a new AI model, a new platform, a new “must-have” tool that promises to revolutionize everything. If you build your curriculum around these, you’re building on quicksand. By the time your students graduate, those tools will be obsolete.
I saw this happen in the early days of mobile app development. Everyone was obsessed with teaching specific frameworks. The developers who only learned one framework were lost when the next one came along. The ones who understood the fundamental principles of mobile UI/UX, data persistence, and network requests—they thrived. They could pick up any new framework in a week.
It’s the same with AI. Stop trying to teach “How to use ChatGPT for X.” Instead, teach the underlying concepts:
- What is a large language model? Not the math, but the intuition. Think of it as a super-powered autocomplete that has read the entire internet. What does that mean for how we get information? How does that change how we write?
- What is a diffusion model? Again, the intuition. It’s a system that learns to create images by starting with noise and slowly refining it into something coherent. What does that mean for art, for design, for creativity?
- What is supervised vs. unsupervised learning? This is a fundamental concept that applies everywhere. Are we teaching the AI with labeled examples, or are we asking it to find patterns on its own? Give students a simple dataset and have them decide which approach is better and why.
Focus on these core ideas. The tools are just temporary implementations of these much bigger concepts. A student who understands the idea of a language model will be able to adapt to any new chatbot that comes along. A student who only knows how to prompt one specific tool will be stuck.
Principle 2: Problems, Not Prompts
The biggest mistake I see in AI education is the obsession with “prompt engineering.” It’s a useful skill, sure, but it’s a tiny piece of a much larger puzzle. It’s like teaching a woodworker how to use a specific brand of sandpaper and calling it a woodworking course. You’re missing the entire point.
Instead of teaching students how to write the perfect prompt, teach them how to solve problems. Give them a real-world challenge and let them figure out how AI can help. The prompt is just one step in that process.
Here’s an example. When I was building RemoteTeam, we had a huge problem with scheduling across different time zones. It was a nightmare. A traditional approach would be to teach students how to use a specific scheduling tool. A slightly better approach would be to teach them how to prompt an AI to generate a schedule.
A much better approach? Give them the problem: “You have a team of 10 people in 8 different time zones. They need to have a 1-hour all-hands meeting once a week. Find the best time for that meeting.”
Now, the students have to think like problem-solvers:
- Deconstruct the problem: What are the constraints? What does “best” mean? (Least disruption, most people in their normal working hours, etc.)
- Gather the data: They’ll need the time zones and working hours for each team member.
- Brainstorm solutions: How could AI help here? Maybe they could use an LLM to parse the data and suggest some initial options. Maybe they could use it to write a small script to analyze all possible time slots.
- Test and iterate: The first few suggestions from the AI might be terrible. The students will have to refine their approach, tweak their prompts, and maybe even combine AI with other tools (like a simple spreadsheet).
In this scenario, the prompt is just a tool, not the goal. The goal is solving the problem. This is how real-world work gets done. No one at a successful startup gets praised for writing a beautiful prompt. They get praised for solving a customer’s problem. By focusing on problems, you teach students how to think strategically, how to be resourceful, and how to use AI as a true partner in the creative process.
Principle 3: Build the “BS” Detector
AI is an incredible tool, but it’s also a world-class liar. It hallucinates facts, makes up sources, and confidently presents complete nonsense as truth. If we are going to unleash this technology on students, we have a moral obligation to teach them how to question it.
An AI curriculum that doesn’t have a strong ethics and critical thinking component is not just incomplete—it’s dangerous. We need to build a healthy skepticism in our students. They need to become expert fact-checkers and develop a finely-tuned “BS” detector.
How do you do this? You actively try to break the AI.
- The Misinformation Challenge: Give students a controversial topic and ask them to use an AI to write a persuasive essay arguing for a specific viewpoint. Then, have them fact-check every single claim the AI made. Where did it get its data? Are the sources real? Is the logic sound? I guarantee you, they will find errors. This is one of the most powerful lessons you can teach.
- The Bias Audit: AI models are trained on human data, and that means they inherit all of our biases. Have students test this. Ask an image generation model to create pictures of “a doctor,” “a CEO,” or “a nurse.” What do the results look like? Are they diverse? Why or why not? This leads to incredibly important conversations about representation, fairness, and the hidden assumptions baked into our technology.
- The “Why” Game: Teach students to constantly ask “why” when interacting with an AI. If the AI gives you an answer, ask it to explain its reasoning. If it suggests a piece of code, ask it to explain what each line does. This forces the student to move beyond passive acceptance and engage with the AI’s output on a much deeper level. It’s the difference between copying an answer and actually understanding it.
I’ve invested in companies like Hugging Face that are at the forefront of responsible AI development. The best minds in the field are all grappling with these ethical questions. To exclude them from an AI curriculum is to give students a dangerously incomplete picture of what this technology is and how it works.
The Real Future of AI Education
So, what does this all mean for the future of your classroom? It means we need to stop thinking of AI as a subject to be taught, and start thinking of it as a tool to be used. It’s a new kind of pencil, a new kind of calculator, a new kind of library. It’s a powerful amplifier for human intellect, but it’s not a replacement for it.
The curriculum of the future won’t have a class called “AI 101.” Instead, AI will be woven into the fabric of every other subject. In history class, students will use AI to debate historical figures. In science class, they’ll use it to design experiments. In art class, they’ll use it to create new forms of media.
This is a huge shift, and it can be intimidating. But it’s also an incredible opportunity. For the first time, we have a tool that can personalize learning for every student, that can automate the boring stuff and free up teachers to do what they do best: inspire, mentor, and guide.
Don’t get caught up in the hype. Don’t worry about which tool is the “best” one. Focus on the timeless skills: problem-solving, critical thinking, and ethical reasoning. If you do that, you won’t just be preparing your students for a world with AI. You’ll be preparing them to build a better one.
Frequently Asked Questions
Do I need technical skills to design an ai curriculum (the counterintuitive guide for educators)?
Not necessarily. While technical understanding helps, the most important skills are clear thinking and the ability to break problems into smaller pieces. Many successful founders I've invested in started with zero technical background and either learned enough to be dangerous or found the right technical partner.
What tools do I need to get started?
Start with the basics. You don't need expensive software or fancy tools. A spreadsheet, a note-taking app, and direct access to your customers will get you further than any enterprise platform. Add tools only when you hit a specific bottleneck.
How do I measure success with this approach?
Pick one or two metrics that directly tie to your goal and track them weekly. Vanity metrics like page views or follower counts rarely matter. Focus on metrics that reflect real engagement or revenue impact.
What are the most common mistakes when designing an ai curriculum (the counterintuitive guide for educators)?
The biggest mistake I see is overcomplicating things early on. Start with the simplest version that works, get real feedback, and iterate from there. Another common trap is copying what worked for someone else without understanding the context behind their decisions.