The biggest AI mistake students are making right now

As artificial intelligence becomes more common in the workplace, students face a new challenge: learning how to use AI without letting it do the thinking for them. Employers increasingly expect graduates to know their way around AI, but they also want problem-solvers who can adapt as the technology evolves.  

Jeremiah Contreras

"If students learn how to use AI to challenge their reasoning and sharpen their judgment instead of replacing it, they'll be miles ahead,” said Jeremiah Contreras, associate teaching professor of accounting at the Leeds School of Business and AI-in-education expert. CU Boulder Today recently spoke with Contreras about the biggest AI mistake students are making, what employers want to see from new graduates and how students can use AI to become better thinkers, not just faster workers.

What’s changed about how students should think about AI this school year?

More businesses have started identifying ways AI can help them operate more effectively. At the same time, they're looking for students who can help find those solutions.

There's now an expectation that students bring curiosity, a basic understanding of AI tools and some level of competency using them.

The biggest change is that those expectations are stronger than they were even a year or two ago. AI has shifted from being a technology with potential to one that's increasingly proving useful in the workplace.

Are employers expecting more from entry-level hires now?

The specific AI tools employers use will keep changing. What they're really looking for is someone who can step back, understand a problem and evaluate which tools or approaches make the most sense for solving it. The solution that works today may not be the solution six months from now. 

I'd rather have a student identify a business problem and develop a solution that isn't perfect than provide the perfect answer to a problem someone else, or AI, identified. The ability to recognize opportunities, define the right problem, and use technology to develop and refine solutions is often more valuable than finding the perfect answer on the first attempt. The problem-solving process is often more important than the final solution.

The key is learning how to think about why a tool solves a problem, not just how to use a specific tool.

What helps students stand out in job interviews?

Employers are looking for proof. They want to see that you've actually done something, not just that you've learned about it.

If a student can say, "I had this problem and built a solution using AI," that tends to be much more compelling than simply saying they took a class.

More employers are asking candidates to solve problems using AI during interviews. They are interested in how candidates approach a problem, use AI appropriately, and explain their thinking. The process often matters more than the final answer. 

What mistakes are students making with AI right now?

Some students take the approach that if they put an assignment into AI, they'll get an answer and earn a good grade.

The biggest mistake is relying too much on AI to do the work.

AI can be great at explaining a concept in a different way or helping you better understand something. But you still need to go back, organize your thoughts and make sure you can explain the material yourself.

What's the right role for AI in the learning process?

Students should use AI to support learning, especially in areas where they already have some familiarity and want another perspective or explanation. When learning something new, textbooks, course materials and other authoritative sources are still important. AI is a great tutor, but it should not replace foundational learning. As students gain expertise, they'll also get better at recognizing when AI gets something wrong.

What does it actually mean to be prepared for an AI-driven world?

From an educational standpoint, we still have to make sure students learn the fundamentals in a world where AI can do many assignments for them. There's a blurry line between using AI to learn and letting it take on too much of the cognitive load.

That's where what I call "productive friction" comes in. AI makes things very efficient. From a business perspective, that's great. But from a learning perspective, students still need to struggle with concepts to truly understand them.

Ultimately, businesses want students who are comfortable with AI. But even more important is the ability to evaluate problems, identify solutions and adapt as technology changes.

How can students use AI in a way that actually makes them smarter, not just faster?

A common approach is to ask AI for an answer, read it and think, "That makes sense."

A better approach is to ask AI to challenge your thinking. Have it ask questions. Have it critique your reasoning. Have it push you to explain your ideas more clearly.

The goal is to use AI as part of the learning process rather than as a shortcut around it.

Can you give an example of using AI well versus using it poorly?

One approach is to upload an assignment, have AI do most of the work and submit the output with very little effort. 

A better approach is for a student to think through the assignment first, develop some ideas and then use AI to challenge or refine that thinking.

What are some practical ways students can build AI skills right now?

Start by identifying a problem in your own life and trying to solve it with technology. Maybe it's comparing colleges. Maybe it's planning meals. Maybe it's managing a schedule or budget.

If students learn how to identify a problem, experiment with tools and build a solution, that translates directly into the kinds of skills employers value.

The more students explore, experiment and stay curious, the better prepared they'll be.

What's an example of a project that would impress an employer?

I had a student build an app to help track personal expenses. Years ago, that would have required a skilled coder. Today, students can create simple solutions much more easily.

It wasn't a business-ready product, but it solved a real problem. Employers love seeing students identify a problem and solve it in an innovative way.

What should students do when schools restrict AI but workplaces expect it?

I think about AI the same way I think about a tutor. I wouldn't want a tutor to do my work for me. I'd want a tutor to explain concepts, challenge my thinking and help me learn.

If a class doesn't allow AI on an assignment, there may be a learning reason behind that policy. Students should respect it.

At the same time, when course policies allow it, they can often use AI outside of the assignment itself to better understand a concept and prepare to do the work on their own.

Do students need to learn specific AI tools?

Not necessarily. It's more important to learn how to think through problems and find the right tool for a given situation.

Curiosity is key. The tools will change. The ability to evaluate a problem, research solutions and adapt is still fundamentally human.

Students should use AI to support learning, especially in areas where they already have some familiarity and want another perspective or explanation. When learning something new, textbooks, course materials and other authoritative sources are still important. AI is a great tutor, but it should not replace foundational learning. As students gain expertise, they'll also get better at recognizing when AI gets something wrong.

Looking ahead, what will separate students who thrive from those who struggle?

The students who thrive will be the ones who stay curious.

There's no single right way to use AI because people learn differently. The key is figuring out how AI can help you learn more deeply rather than simply work faster.

Students who learn how to do that will be more effective, more adaptable and better prepared for whatever comes next.