I predicted layoffs and shockwaves: Here’s what we’ve learned from a year of AI experiments

Exactly one year ago, I stood on stage at the Digital Tech Summit in Øksnehallen and stated – somewhat boldly – that 2025 would be the year when generative AI would genuinely start reshaping the Danish labour market. I said we might see redundancies or significant shifts in staffing, because AI would take over a substantial part of employees’ tasks. In other words: robots would change how we work.

Now, as the year draws to a close, I must acknowledge that I was wrong. Not about what the technology can do – we are only at the beginning of that journey – but about the speed of adoption. While generative AI has become mainstream, embedding it in the core business is taking longer than expected.

The large-scale layoffs have, fortunately, not materialised. We are seeing fewer student roles and fewer junior hires, but core teams remain in place. Many people now use AI in their daily work – Copilot, Gemini, ChatGPT and new AI-powered browsers such as Atlas are now part of meetings, emails, research and reporting. That has increased efficiency. But the clear, measurable bottom-line impact is still emerging.

Why is it so difficult to move beyond pilots?

The barriers are not in the technology itself, but in the complex organisational ecosystems into which it has to integrate: legacy IT, uneven data quality, limited skills, regulatory complexity, organisational friction and deeply embedded habits. It is not difficult to generate long lists of promising use cases, nor to run pilots. The real challenge is moving from pilot to production. Many organisations remain stuck at that point.

At the same time, generative AI introduces new complexity compared to traditional AI. Because it can generate new content – write, design, code or act as an assistant – it is harder to get governance right. This makes it particularly challenging in regulated sectors such as banking, insurance, public administration and pharmaceuticals. Responsible and ethical AI use is therefore not just a political priority; it is a prerequisite for safe deployment.

AI is changing who we hire – not necessarily how many

Emerging patterns show that generative AI shifts roles, rather than simply reducing headcount. Many routine analytical and writing tasks can now be performed more efficiently by experienced employees with AI at hand. This reduces demand for junior positions – but increases the need for talent who can apply and operationalise AI.

Some sectors are already deploying generative AI at scale, particularly where the value chain is digital. In media, news copy and image suggestions are produced with AI as co-author. In tech, AI automates coding and quality assurance. Microsoft CEO Satya Nadella has stated that up to 30% of the company’s code is now written by AI.

The coming year will be critical

Are Danish organisations falling behind? And is the AI bubble about to burst? No to the first – and possibly no to the second. The coming year will likely be less about experimentation and more about integration, standardisation and governance. And even if we are in a bubble, we should remember: the internet didn’t disappear when the dotcom bubble burst.

Read the article in Danish in Version2.