Lessons From A Year Of Teaching An AI Course
I've spent the last year teaching AI to people who never asked to learn it. Not developers. Not data scientists. Regular adults in their 50s, 60s, and 70s who keep hearing about ChatGPT and want to know what the fuss is about. We built an AI education platform to answer that question, and along the way, we learned something that changed how we design everything: what beginners want to know is nothing like what most courses teach.
The Questions Nobody Expects
We collected the most common questions from our first 500 learners. The top five were:
- Is it safe to use?
- Can it see my personal information?
- Is it free?
- What would I actually use it for?
- Am I going to break something?
Notice what's missing. Nobody asked about prompt engineering. Nobody wanted to understand Large Language Models. Nobody cared about the difference between GPT-4 and Claude. They wanted reassurance first, practical use cases second, and technical detail a distant third.
Most AI courses lead with the technical detail.
Why "Start With The Basics" Fails
The standard approach to AI education follows a familiar pattern: define AI, explain Machine Learning, introduce neural networks, then move on to applications. It mirrors how computer science departments have taught for decades.
For a retired teacher who wants to know if ChatGPT can help her write a speech for her friend's 70th birthday, this approach is like explaining combustion engines before letting someone drive a car. The "basics" that courses assume are the basics aren't basic at all. They're foundational concepts from a technical discipline. That's a different thing entirely.
The actual basics for a non-technical adult are:
- Where do I go to use this? (A website. You don't install anything.)
- Do I need to pay? (No. The free versions are good enough to start.)
- What do I type? (Whatever you'd say to a helpful person.)
- Will it remember what I told it? (Not between conversations, usually.)
These four answers take about two minutes. After that, you can hand someone a laptop, and they'll figure out the rest through doing. But most courses won't get to this point until module three.
What Completion Rates Actually Tell Us
We track completion rates across our 10-module course. The pattern is consistent: modules that open with a practical task ("Try asking ChatGPT to write a thank-you note") have 85–90% completion. Modules that open with explanation ("In this section, we'll look at how AI generates text") drop to around 60%.
Same learners. Same platform. Same week. The only difference is whether we asked them to do something or understand something first.
This isn't a criticism of theory. Understanding how AI works matters, especially around safety and privacy. But the sequence matters more than the content. Do first, understand later. That's how most adults learn new technology anyway. You didn't read the iPhone manual before making your first call.
The Privacy Question Is The Real Gatekeeper
Every AI course I've reviewed treats privacy as a footnote. A section near the end, after the exciting use cases. Usually, a paragraph that says "Be careful what you share" and moves on. For learners, privacy isn't a footnote. It's the main event. Over 40% of our new users cite privacy concerns as the reason they haven't tried AI yet. Not lack of interest. Not lack of access. Fear.
When we moved our privacy and safety module from position 8 to position 2, overall course completion jumped by 23%. People weren't dropping out because the content was hard. They were dropping out because they didn't trust the tool they were being asked to use.
Address the fear first. Everything else gets easier.
Designing For The Real Learner
After a year of iterating, here's what works for adult AI learners who aren't from a technical background:
- Lead with action, not theory. The first thing a learner should do in any AI course is type something into ChatGPT and get a response. Not read about it. Not watch a video about it. Do it. That single moment of "oh, that's what it does" is worth more than any explanation.
- Answer the safety question early and honestly. Don't wave it away. Don't bury it. Show them exactly what data is shared, what isn't, and how to use the tool without exposing anything personal. Be specific. "Don't type your bank details" is more useful than "be mindful of your digital footprint."
- Use their vocabulary, not yours. If your course materials include the word "parameters," you've already lost a chunk of your audience. This isn't about dumbing things down. It's about meeting people where they are. "Settings you can change" works just as well and doesn't make anyone feel silly.
- Give them a reason that matters to them. "AI can increase productivity by 40%" means nothing to someone who's retired. "You can ask it to explain your prescription side effects in plain English" means everything. Know your audience. Pick examples from their world, not yours.
- Build confidence through small wins. Our most successful module asks learners to use ChatGPT for three things in one week: plan a meal, draft a short email, and find out something they've always been curious about. By the end of the week, they're no longer beginners. They're users.
The Opportunity Most AI Course Designers Are Missing
There are roughly 20 million adults over 55 in the UK alone. Most of them have smartphones, home broadband, and an active curiosity about AI. They read about it in newspapers. They hear about it from grandchildren. They want to understand it.
The market for accessible, non-technical AI education is enormous, and it's being almost entirely ignored by the eLearning industry. The courses that exist are built for professionals, career changers, and students. The fastest-growing demographic of new internet users is being left to figure it out from YouTube videos and newspaper columns.
That's the gap. And it won't stay open forever.
What I'd Ask Every Course Designer To Consider
Before you publish your next AI course, try this: give it to someone over 55 who's never used ChatGPT. Don't help them. Don't explain anything that's not in the course. Just watch.
If they get confused in the first five minutes, your course has a sequencing problem. If they ask "But is it safe?" before you've covered it, you've buried the lead. If they close the laptop and say, "I'll come back to it later," they probably won't.
The technology isn't the barrier. The teaching is.
