Most people still have no idea how to use AI tools effectively. That's the core takeaway from a Guardian piece by a writer who has spent significant time teaching non-technical people how to work with AI, published March 10.
This tracks with what we see constantly in the AI productivity space. There's a massive gap between the people building these tools, the early adopters who live on Twitter talking about "prompt engineering," and the vast majority of workers who opened ChatGPT once, typed a vague question, got a mediocre answer, and never came back.
The problem isn't the tools. ChatGPT, Claude, Gemini - they're all genuinely useful for real work. The problem is that using them well requires a specific kind of skill that nobody teaches. You need to know how to break tasks down, give context, iterate on outputs, and recognize when the AI is confidently wrong. None of that is obvious the first time you open a chat window.
The Teaching Problem No One Talks About
AI companies spend billions on model capabilities and almost nothing on user education. OpenAI's help docs read like they were written for developers. Anthropic's prompt guides assume you already know what you're doing. Google just throws features at the wall.
Meanwhile, the people who would benefit most from these tools - small business owners trying to draft proposals, teachers building lesson plans, freelancers managing client communications - are left to figure things out alone or rely on breathless YouTube tutorials that promise "10x productivity" without teaching any transferable skills.
The real unlock for AI adoption isn't better models. It's better teaching. And right now, that teaching is happening mostly through informal channels: coworkers showing each other tricks, blog posts from practitioners, and articles like this Guardian piece.
Until AI companies take user education as seriously as they take benchmark scores, the gap between what these tools can do and what most people actually do with them will keep growing.