The Craft Crisis: Are We Automating Away the Next Generation of Experts?

Executive-Summary: The 2026 entry-level hiring crisis is being fuelled by an “AI Expertise Gap.” Marketing and tech veteran explains why juniors are losing the chance to learn the craft and proposes a new statutory probation strategy to save the future workforce.

Two weeks ago, I went down to Southampton for a teaching session at the college. While I was there, I took the opportunity to catch up with some friends. As usually happens whenever I’m involved, the conversation eventually turned to AI—but this time, what I heard gave me a real reason to pause.

My friend is a software engineering leader for one of the top international developers in the UK. He was praising the perks of AI, marveling at how fast it makes the build process. But then he hit me with the “but.” He explained that he spends a massive amount of his time reviewing every single line of code the tool generates to ensure it’s actually correct.

Then he asked the “billion-pound” question: “I can do this because I have 10 years of experience writing code from scratch. But what happens to the graduates? If AI writes the code for them from day one, when do they ever get the chance to learn the craft?”

Why “Old School” Theory Still Matters

My answer to him was that this is where education must pivot. Students need to master the foundations from scratch before they are ever allowed to touch an LLM. But let’s be real—is that actually happening?

It took me back to my days studying journalism in 2012. At uni, I used to complain constantly about the “ancient” systems they forced us to learn—old-school layout and typesetting (enmaquetado) that felt like a waste of time because nobody in the industry used them anymore.

But I was wrong. The theory of doing things “the old way” gave me a mental map. It meant that when the tech evolved, I understood the why behind the how. I adapted faster because I knew the foundation.

My friend in Southampton isn’t alone. I’ve spoken to other developers since, and they all have the same fear: We are building a generation of “Prompt Engineers” who don’t actually know how to build anything.

The Business Risk: Catastrophe in the Making

If you’re a business owner, you might be wondering why this is your problem. It’s simple: If you hire juniors who know how to use the technology but don’t know if the output is right or wrong, the results for your company will be catastrophic.

And the data for 2026 shows the storm is already here.

Earlier this year at Davos, Google DeepMind CEO Demis Hassabis admitted that AI is starting to hit the junior level hard. He isn’t the only one sounding the alarm. Anthropic’s Dario Amodei previously warned that AI could wipe out half of all entry-level white-collar jobs. Now, Microsoft AI CEO Mustafa Suleyman is telling the Financial Times that tasks for lawyers, accountants, and marketers will be “fully automated” within the next 12 to 18 months.

The 2026 Entry-Level Crisis

The data from America proves this isn’t just a “software” problem. The entry-level hiring crisis is peaking. More than half of employers surveyed by NACE rated the 2026 job market as “poor or fair“, with a measly 1.6% increase in hiring for this year’s graduating class.

We are facing a double-edged sword:

1. The Opportunity Gap: A lack of roles for graduates due to a struggling global economy.
2. The Skills Gap: Graduates who aren’t prepared for a world where they are expected to “supervise” AI rather than just “use” it.

My Challenge: A New Probation Strategy

So, what do we do? We can’t just sit back and watch the “craft” die.

Having an AI strategy isn’t “nice to have”—it’s a survival requirement. This strategy must include a completely new way of onboarding. It is no longer enough to have a standard 3 or 6-month probation where someone just learns “how we do things here.”

We need to slow down to go fast.

Employers must dedicate the first phase of employment to training and supervising new hires to complete tasks without the help of AI. We need to see they can do the work from scratch. Only then should we show them how to incorporate AI to scale that work.

Yes, the process is slower. Yes, it’s more expensive. But at a time when the current workforce is burning out, we need new talent more than ever. This is the only way to ensure we aren’t just hiring ‘button-pushers’ who will eventually lead our companies into a wall. It’s time to stop looking for shortcuts and start protecting the craft.