AI Is Changing How Developers Learn Code, But Is It Creating A Confidence Gap?

AI coding tools have rapidly become part of the modern developer toolkit. Whether it’s generating boilerplate code, debugging issues, or explaining complex concepts, platforms such as GitHub Copilot and ChatGPT are changing how software is built and how aspiring engineers learn their craft.

But as a new generation of developers enters the workforce having learned to code alongside AI, employers are beginning to ask a different question. It’s no longer simply whether junior developers can use AI, but whether they have developed the judgment to use it effectively.

New research from BairesDev’s quarterly Dev Barometer, which surveyed 1,569 developers across 77 countries, suggests that there may be a disconnect between how junior developers view their abilities and how more experienced engineers assess them.

And as graduation season brings a fresh wave of talent into the job market, the findings offer an interesting glimpse into how AI could be reshaping software development careers.

AI Is Becoming A Core Part Of Learning

According to the survey, 85% of junior developers say AI improves their understanding of software development. For many early-career engineers, AI is no longer just a productivity tool. It’s a tutor, mentor and coding assistant rolled into one.

Now, this isn’t necessarily a bad thing. Historically, developers have relied on documentation, forums, Stack Overflow and colleagues to learn new concepts. AI simply provides another way to access information and guidance, often in real time.

The result is that junior developers can move faster, experiment more freely and potentially learn concepts that might otherwise have taken longer to grasp.

But the important thing to remember is that speed and understanding aren’t always necessarily the same thing.

The Confidence Gap

Perhaps the most striking finding from the report is that only 16% of senior developers believe junior engineers fully understand the code they generate using AI tools.

That statistic highlights what could become one of the defining workplace tensions of the AI era.

Senior engineers appear to be asking a different question from their junior colleagues. While juniors may focus on whether they can successfully complete a task, experienced developers are often more concerned with whether someone understands why the code works, how it can fail and how it should be maintained over time.

The difference between the concerns of junior and senior engineers is important, because software development is rarely just about producing code. Rather, it’s about creating systems that can scale, evolve and remain reliable years after they are first deployed.

The concern isn’t necessarily that AI generates poor code. Rather, it’s that developers may become increasingly capable of producing working solutions without fully understanding why those solutions work, where they might fail or how they should evolve over time.

As AI takes on more routine coding tasks, employers appear to be placing even greater value on reasoning, technical judgment and systems thinking from the very start of a developer’s career.

Can Developers Still Code Without AI?

The survey also found that nearly one in four junior developers admit they’re not confident writing code from scratch without AI assistance.

At first glance, that statistic may sound concerning. However, it also reflects a broader shift taking place across knowledge work. Accountants use software, designers use AI-powered creative tools, marketers use automation platforms and now developers are increasingly using AI to code.

The real question may not be whether developers can work without AI, but whether they can recognise when AI is wrong. As AI-generated code becomes more sophisticated, the ability to review, validate and challenge outputs will become just as important as the ability to write every line manually.

What Skills Matter Most Now?

Interestingly, junior developers don’t actually seem to believe that coding itself is the primary route to employment. In fact, only 5% of respondents said writing clean code is the most important skill for getting hired today.
Instead, almost half (48%) ranked problem-solving and analytical thinking as the most valuable skill, while 18% pointed to AI tool proficiency.

This suggests that younger developers may already recognise that employers are increasingly hiring for adaptability, reasoning and business understanding rather than simply technical execution.

In many ways, that reflects the reality of modern software development. AI can generate code, but it can’t always identify the right problem to solve, understand customer needs or make strategic decisions about product development.

Those are things that remain distinctly human skills, at least for the time being.

The survey also suggests employers are increasingly looking beyond technical knowledge alone. Senior developers ranked real-world project experience, internships and practical coding exercises as the strongest indicators that junior engineers are ready to contribute. In other words, evidence that candidates can apply what they’ve learned is becoming just as important as the knowledge itself.

A New Era For Software Development

The findings arrive at a time when the software industry is still working out what AI-native development looks like.

For decades, learning to code involved mastering syntax, understanding algorithms and building projects line by line. Today’s graduates are entering a world where AI can perform many of those tasks instantly.

That doesn’t mean coding fundamentals have become less important. If anything, AI is increasing the value of understanding software architecture, security, debugging and systems design because developers must now evaluate and improve AI-generated outputs rather than simply produce them.

The challenge for educators, employers and developers themselves today, however, is finding the right balance.

AI is clearly becoming an essential part of software development. The question is whether the next generation of engineers will use it to deepen their understanding or whether reliance on automation will gradually replace it.

The answer could shape not only how developers are hired, but what being a developer actually means in the years ahead.