Never Stop Wondering: Why Curiosity Is the Missing Skill in AI Upskilling

Never Stop Wondering

I vividly remember my Year 8 Chemistry lesson on reactive metals. Not for the thrill of watching the TV being wheeled into class or for seeing the violent spectacle when a small piece of metal hit water, but for a different kind of volatile reaction altogether. We'd been working down the periodic table, watching the reactions get more dramatic for each element. I spotted the pattern, and asked the teacher...“Could we use elements like this as a weapon, dropped in the sea?”.

Instead of engaging with the thinking behind it and using it as a chance to explain why it was impractical and very dangerous, the teacher just dismissed it outright as stupid. That was the last time I put my hand up at school. From then on, I did all my wondering in my head. This was long before I had access to the internet, so answers weren't always easy to find.  Sometimes it took years, a random magazine or documentary to finally stumble on one. Losing the ability to figure things out on the fly with others stops everyone nudging towards the answer faster. Scale that across a classroom, team or workforce and you lose a form of information exchange.

Rapidly accelerating technologies like AI promise great transformation, but they also stir great fear. We need more minds on this problem, even the uninitiated to get this right. We need novices asking novel questions, helping all of us figuring out where we should be using this technology, what its downsides are, and how we should use it in additive ways to our own skills. Inquisitive questions often reveal gaps in other digital skills and remind us that we're all at different starting points, even if we've had the same exposure. Upskilling is never one size fits all, nor is it linear.

What I actually see

I've spent years helping people find better ways of using technology, data and AI in their work. Vitally, I learn from them and their questions too. I've seen this first-hand running employer sessions with TechSkills, where real workplace problems get brought straight into the classroom. It creates exactly the kind of space where curiosity is encouraged and no question is a stupid one. A question can challenge my view of what’s possible, improve my approach to training or suddenly get everyone in the room on the same level of understanding better than any pre prepared example can.

Some learners naturally run ahead. Confident enough to see an example and bend it to their domain or interests without even asking for permission. Others struggle with the abstraction of an example that isn’t perfectly tailored to their situation. This seems particularly prevalent with tech skills.

People are strangely open, and exceedingly harsh, about their own technical skills. "I'm not a technical person" has become a protective shield, lowering expectations of what someone thinks they are capable of. You’ll never hear people as openly admit to poor communication skills, or to not being very good at critical thinking. That mindset stops people from wondering, from experimenting, from feeling confident enough to try and acquire a new skill.

Then there's the language around tech. Prompt engineering is just about the most unapproachable term for interacting with the most approachable technology. I'm not saying we need to dumb down, we just need to clear the barriers that are contributing to a lack of confidence. Many already know how to tackle a problem with AI but still seek permission and reassurance before attempting it. More than once, people have reached out to me with a well-written message asking how they should prompt AI to do a given task. Their message contained every element of a good prompt. They'd broken down the task, thought through and described what the output should look like, shared extra context and even selected the right AI tool. But they couldn't take that last step, copy and paste the message into the tool themselves. On one occasion, this was a full year after someone had received training. A year of thinking they weren't ready when they always were. Regardless of the new skill you want to learn, you will likely never feel fully ready and put it off until an imaginary future where your calendar is miraculously clear. Start small, start today. To borrow a term from the stock market, it’s “Time in the market, not timing the market."

When it comes to applying the nascent skill, another issue is patience. People feel the need for immediate results perhaps to feel they have caught up or worse to relieve a certain work pressure they face. However, a task too big, too complex can erode confidence. Worryingly, many go straight to the most difficult task, the one they add real value to currently. They don't consider whether AI should even tackle this, or what they lose if it succeeds. They skip the easier learning exercises that would build their confidence and expertise gradually. When the output then isn't right, they don't always consider how they can get better at this, or what they could try differently. Instead, they conclude either they're not good enough, or the tool isn't good enough.

AI isn’t magic in the same way cars aren’t or the internet isn’t. But how many of us have actively wondered how everyday technologies work? Most people can’t draw a mechanically accurate bicycle from memory (try it yourself). We don’t need to know every facet inside out but a sense of curiosity and deeper understanding of how things work will help us to know how and when to use them better.

How can we change this?

We all must take responsibility and the agency to be curious and act on our ideas, even when it feels uncomfortable and even if they don’t go particularly far. We should pattern match, spotting an example technology and workflow in one domain and swap out elements to make it relevant for our world and experiment.

We need to reconsider education systems built to reward conformity and set answers over experimentation and novel approaches. We need to challenge workplaces with a “just the way we do things around here” mentality and encourage people to rethink workflows and work itself from the bottom up.

Any process or way of doing something that's gone unchanged for even six months is probably outdated. There's likely a better approach, better technology, more information, more skills and tools to call on. Constant change and perfect optimization are not the goal but neither is standing still. Ideas come when you expand your horizons. Conversations across teams, roles, and sectors surface answers you'd never find alone.

Making a workforce and future generations more confident and curious can only be a benefit in a rapidly changing world with an aging population. Which means fixing it is a group effort. Educators, Employers, and anyone designing a pathway into a digital career all need to encourage the role of curiosity in learning and work. Through my work with  Techskills, I've seen how much learners gain when employers bring real problems into the classroom. It creates exactly the kind of space where curiosity is encouraged.

In a world where automation and productivity seem to be the main drivers, I’m advocating for a slower approach where people have more agency to be curious and figure out how technology can add value to them and their work. I'm certain a curious, confident, engaged, upskilled workforce becomes a more productive one. I'm not sure focusing solely on productivity gets you those other benefits.

Daniel runs Galagog combining training, research, and advisory to help organizations build AI literacy, digital fluency, and the human skills needed to make sense of tomorrow and thrive in a changing world.