Despite the whirlwind of hype, new research suggests that generative AI chatbots like ChatGPT and Claude have, so far, made barely a ripple in the labour market, leaving jobs and wages largely untouched.
A Grounded Reality Check
A comprehensive study by economists Anders Humlum (University of Chicago) and Emilie Vestergaard (University of Copenhagen) has found that the economic impact of generative AI chatbots on workers has been negligible. Their paper, Large Language Models, Small Labor Market Effects, analysed data from over 25,000 workers across 7,000 Danish workplaces between 2023 and 2024, focused on 11 occupations considered highly susceptible to AI disruption, including software developers, journalists, legal professionals, and teachers.
No Significant Impact
According to findings published in the paper, despite widespread adoption, most firms encouraged chatbot use, 38 per cent deployed in-house models, and 30 per cent of employees received AI training. Crucially, the study found no significant changes in earnings or working hours across any occupation. The researchers, therefore, concluded that “AI chatbots have had no significant impact on earnings or recorded hours in any occupation.”
Just Modest Productivity Gains
It seems that while users reported modest productivity gains, averaging time savings of 2.8 per cent to 5.4 per cent of their weekly hours, these did not translate into reduced workloads. In fact, AI adoption led to new tasks for 8.4 per cent of workers, such as supervising AI outputs or adapting workflows to accommodate the technology.
Other Research Supporting The Findings
Other recent research has also reached similar conclusions. For example, a separate analysis by Barclays, led by economist Mark Cus Babic, examined AI exposure across various occupations and industries in the U.S. and Europe. The study found that less than 10 per cent of core job tasks could be better performed by AI. Interestingly, occupations most exposed to AI were not always at risk of being replaced. For example, while roles like proofreaders and typists are more susceptible to automation, professions requiring significant interpersonal skills, such as translators, are less replaceable.
Contrary to fears of widespread job losses, therefore, the Barclays analysis found that AI exposure correlated with employment growth, not reduction.
However, this study also noted that AI exposure was linked to slower wage growth, with rising AI exposure reducing annualised wage growth by up to 0.74 percentage points.
Contrasting Findings
While these studies suggest a limited immediate impact of generative AI on jobs and wages, other recent research presents a more nuanced picture. For example:
– A 2024 PwC report found that sectors with high AI penetration experienced nearly fivefold greater labour productivity growth compared to less exposed sectors. In the UK, job postings requiring AI skills were growing significantly faster, with employers offering a 14 per cent wage premium, particularly in legal and IT roles. The findings were based on global employment and productivity data tracked by PwC’s Economic Outlook research team.
– A 2023 study by researchers from the University of Oxford and the University of Copenhagen, analysing online labour market data from platforms like Upwork and Freelancer, observed a decline in demand for text-related and programming-related jobs following the introduction of ChatGPT. However, the remaining jobs in these submarkets became more complex, and competition among freelancers increased, suggesting a shift in the nature rather than the volume of work.
– Joint research published in 2023 by the International Labour Organisation and the World Bank indicated that generative AI could potentially automate between 2 per cent and 5 per cent of jobs across Latin America and the Caribbean. The study warned that women and younger workers in formal employment sectors were likely to be disproportionately affected, especially in roles involving routine cognitive tasks.
The Implications
In terms of the implications of the most recent University of Copenhagen research, the minimal immediate impact on jobs and wages may prompt AI developers to reassess their value propositions. It seems that while the technology holds promise for enhancing productivity, the anticipated economic benefits have yet to materialise at scale.
Also, based on these findings, companies investing in AI may want to temper expectations regarding short-term labour cost savings. Instead, the focus could shift towards leveraging AI for incremental efficiency gains and exploring new business models that integrate AI capabilities.
In terms of what this could mean for governments and policymakers, the findings appear to suggest that fears of an imminent AI-induced employment crisis may be overstated. However, the potential for AI to reshape job tasks and create new roles underscores the need for policies that support workforce adaptability, such as reskilling initiatives and education reforms.
As for workers, while it seems (according to this study) that AI has not yet led to significant job displacement, its integration into the workplace is undoubtedly altering job responsibilities. This could mean that workers may need to adapt by acquiring new skills and embracing lifelong learning to remain competitive in an evolving job market.
Perception vs Reality?
One of the more striking contrasts emerging from this research is the growing gulf between how AI is perceived and how it’s actually performing in economic terms. It seems that public debate has largely centred around the threat of mass job displacement, with headlines warning of “white-collar extinction events” and sweeping automation of knowledge work. Yet the data so far simply doesn’t back that up.
For example, a 2024 Ipsos MORI survey found that 61 per cent of UK workers believe AI will significantly reduce job availability within the next decade. However, this fear appears to be driven more by speculation and media narratives than current evidence. Researchers like Humlum and Vestergaard stress that even in sectors with widespread chatbot adoption, measurable labour impacts have been “remarkably muted”.
This mismatch between expectations and evidence could have real consequences, potentially fuelling anxiety, political pressure, or misaligned policy responses. It also raises a challenge for AI companies and advocates, i.e. how to communicate realistic use cases and limitations without losing investor interest or public trust.
What Does This Mean For Your Business?
What this research ultimately seems to reveal is a still-unfolding story, and one that is far less dramatic than the early hype may have suggested. While it’s true that generative AI is being widely adopted across white-collar industries, it looks as though the impact on wages and jobs appears, for now, to be largely neutral. That’s not to say AI isn’t changing the workplace (far from it). However, the kind of sweeping disruption that many predicted simply hasn’t (yet) arrived.
For UK businesses, this latest research provides a valuable window of clarity. It means that rather than expecting AI to deliver immediate cost savings through workforce reductions, firms may find more tangible returns in using chatbots to refine workflows, support staff with repetitive tasks, and free up time for more valuable work. In practical terms, that means revisiting where AI fits in the broader business model, not as a silver bullet for efficiency, but as a support tool, and one that still needs oversight, training, and adaptation to work effectively.
For governments, the findings highlight the importance of measured, evidence-based policymaking. While it’s right to prepare for potential shifts in the labour market, it seems there’s currently no need for panic. The real focus might be better placed on supporting agility within the workforce, e.g. through investment in digital skills, better access to lifelong learning, and guidance for employers on effective technology adoption.
Meanwhile, for AI developers, the study is a reminder that user adoption doesn’t always equal economic impact. The technology may be advancing rapidly, but converting that into broad-based value remains a work in progress. As such, the next wave of innovation may need to focus less on scaling up infrastructure, and more on proving real-world outcomes, especially for sectors still unsure how to integrate these tools meaningfully.
In short, this research invites recalibration and puts things a little more in perspective. Generative AI is here, it’s being used, and it’s shaping how work gets done, yet its impact (at least for now) appears to be evolutionary rather than revolutionary. The real question may no longer be whether AI will replace jobs, but whether we’re ready to redesign the way we work alongside it.
