Not things about AI. Things you've always wanted to understand better. A language you've been meaning to pick up. A professional skill you know you should develop. A subject you were curious about at school but never had time to pursue. The history of something you care about. The theory behind something you do every day at work but have never properly understood.
This is one of the most underappreciated uses of AI, and it doesn't get nearly as much attention as image generation or chatbot conversations. When you use a chatbot as a learning partner rather than a question-answering machine, something different happens. Instead of typing a question and getting a response, you enter into a back-and-forth where the AI can explain a concept, check whether you've understood it, offer an analogy when you haven't, give you practice questions, adjust its explanations based on what you already know, and keep going for as long as you want to keep learning.
It's not a replacement for a human teacher. It can't read your facial expressions, notice when you're confused but too polite to say so, or draw on years of experience working with learners in your specific context. But it has some real advantages too: infinite patience, instant availability, no judgement when you ask something basic, and the ability to explain the same concept seventeen different ways until one of them clicks. For self-directed learning (the kind of thing professionals do when they want to develop a new skill or understand a new topic) it's surprisingly useful.
This Thing asks you to experience AI as a learning tool firsthand. You'll use a chatbot as a personalised tutor, try a purpose-built AI learning tool, and reflect on how these compare to the ways you normally learn.
How chatbots work as learning tools
You already know how to have a conversation with a chatbot; you've been doing it since Thing 2. Using one for learning isn't fundamentally different, but the way you approach the conversation changes. Instead of asking a question and accepting the answer, you're trying to build understanding over time.
The key is to treat the conversation as a dialogue, not a search. Here's what that looks like in practice.
Start by setting the context. Tell the chatbot what you want to learn and where you're starting from. "I want to understand how the UK pension system works. I know almost nothing about it beyond the fact that I have a workplace pension" gives the AI much more to work with than "explain pensions." The more honest you are about your starting point, the better the explanation will be pitched.
Ask it to explain, then check your understanding. After the AI explains something, try restating it in your own words and asking whether you've got it right. This is a learning technique called retrieval practice, and it works whether your conversation partner is human or artificial. Something like "So if I'm understanding this correctly, the state pension is based on National Insurance contributions, and you need 35 qualifying years for the full amount. Is that right?" forces you to process what you've just read rather than skimming past it.
Request different angles. If an explanation isn't landing, ask for an analogy, a concrete example, or a simpler version. "Can you explain that using an everyday analogy?" or "Give me a concrete example of how this would work for someone earning £30,000" often unlocks understanding in a way that a more abstract explanation doesn't.
Ask it to quiz you. This is where things get properly useful. Ask the chatbot to give you five questions on what you've just covered, then try to answer them before looking at the AI's answers. You can ask for multiple-choice questions, open-ended questions, or scenario-based questions, whatever suits how you learn.
Build on what you've covered. As the conversation progresses, reference earlier parts. "You mentioned earlier that there's a difference between defined benefit and defined contribution pensions; can we go deeper into that?" This kind of threading is something chatbots handle well, and it mirrors how a good tutorial conversation works.
Be honest when you're lost. There's no social pressure in a chatbot conversation. If something doesn't make sense, say so. "I'm completely lost. Can you start again from the basics?" works perfectly well and costs you nothing.
What chatbots are good at (and not good at) for learning
The strengths are real. Chatbots can explain concepts at whatever level you need, generate unlimited practice questions, provide instant feedback, work through problems step by step, and adapt the pace entirely to you. For factual topics where there's a clear body of knowledge to understand (how a regulatory framework works, the basics of a programming language, the principles of project management, the grammar of a new language) they're genuinely effective.
The limitations are equally real. Chatbots can get things wrong, especially on niche or rapidly changing topics. They can't assess whether you've truly understood something in the way a human tutor can; they can only check whether your stated understanding matches the correct answer. They don't know when you need encouragement versus when you need to be challenged. And they lack the real-world experience that makes a human expert's explanations come alive with stories, context, and hard-won wisdom.
The practical implication is straightforward: chatbot tutoring works best for building foundational understanding of a topic, but you should verify important claims (especially statistics, dates, and niche details) and seek out human expertise for the nuanced, experiential knowledge that AI can't provide. Everything you learned in Thing 15 about hallucinations applies here too, perhaps even more so, because when you're learning something new you're least equipped to spot when the AI has got something wrong.
Purpose-built AI learning tools
Beyond using general chatbots as tutors, several tools have been designed specifically for AI-assisted learning. These are worth knowing about because they've been built around learning science in ways that a general chatbot hasn't.
The learning science behind it
There's a reason these approaches work, and it's not just because AI is impressive. Several well-established learning principles are at play when you use AI as a tutor.
Active recall is testing yourself on material rather than just re-reading it, and it's one of the most effective learning strategies research has identified. When you ask a chatbot to quiz you and try to answer from memory, you're using active recall.
Spaced repetition means revisiting material at increasing intervals. It's another evidence-based technique. You can do this manually by returning to a chatbot conversation a few days later and asking it to quiz you on what you covered previously.
The Socratic method is learning through guided questioning rather than being told answers. This is what Khanmigo is explicitly built around, and you can prompt any chatbot to work this way: "Don't give me the answer directly. Ask me questions that will help me figure it out."
Elaborative interrogation means asking "why" and "how" questions rather than just "what." When a chatbot explains something, asking "why does that work?" or "what would happen if this were different?" pushes your learning beyond the surface.
None of these techniques are new. They've been used by human teachers for centuries. What AI adds is availability: you can use these techniques at any time, at your own pace, on any topic, without needing to find a willing tutor or join a class. That's particularly valuable for professionals who need to learn continuously but don't always have access to formal training.
Resources to explore
Free tier available. Works well as a general-purpose tutor for most topics.
Free tier available. Particularly good at detailed, nuanced explanations and patient back-and-forth.
Free tier available. Integrates with Google services if you want to connect learning to your existing tools.
Free with ads; Duolingo Max (AI features) from approximately £10/month. Available on iOS, Android, and web.
Free for personal use. Best for maths, science, and programming topics.
Free tier with limited AI features. Q-Chat available for conversational study.
Free with a Google account. Upload study materials for grounded AI tutoring and audio overviews.
Evidence-based resources on why testing yourself works, if you want to understand the learning science further.
Activity: your AI learning session
This activity asks you to spend 30 minutes using AI as a personalised tutor, then reflect on how the experience compares to other ways you learn. You'll need a chatbot (any of the free options you've used during this programme) and, optionally, one of the purpose-built learning tools.
-
Choose your topic. Pick something you genuinely want to learn about. The activity works much better if you're actually curious about the subject, because you'll be more engaged in the conversation and better able to judge how well the AI is teaching you.
-
Set up your learning conversation. Open a chatbot (ChatGPT, Claude, or Gemini all work well) and start the conversation by telling the AI what you want to learn, what you already know, and how you'd like the session to work.
-
Have your 30-minute learning session. Set a timer for 30 minutes and engage properly with the conversation. Don't just read the AI's explanations passively. Respond to them, answer quiz questions before checking, restate concepts in your own words, and request different explanations when one doesn't land.
-
Try a purpose-built tool (optional but recommended). If time allows, spend 10–15 minutes trying one of the purpose-built learning tools mentioned in the resources above.
-
Write your learning journal. Compile your reflections into a learning journal entry covering the prompts below.
Why this matters
Professional development is a constant for most working people. There's always something new to learn, a skill to develop, a topic to get to grips with. Formal training is expensive, time-consuming, and not always available when you need it. Books and articles are accessible but passive. YouTube is useful but unstructured.
AI tutoring doesn't replace any of these, but it fills a specific gap. It gives you an on-demand, patient, adaptable learning partner that's available whenever you have 30 minutes and a question. It's particularly good for building foundational understanding of a new topic, revising material you've covered before, and exploring areas where you don't yet know enough to ask the right questions of a human expert.
The experience of actually using AI to learn something, rather than just reading about the possibility, is what makes this Thing valuable. Once you've done it, you know whether it works for you. And if it does, you've added a genuinely useful tool to how you approach your own development.
Claim your Open Badge
Submit your learning journal entry with the topic you studied, the prompts you used, your summary of what you learned, your comparison of AI tutoring with other learning methods, and your reflection on whether you'd use this approach going forward.
Submit your learning journal and reflection to claim this badge via cred.scot.
Claim now