Over the past decade, automation technologies have been a key driver of change in the labour market. There has been a huge increase in the demand for AI skills, which has led to different recruitment processes, different job titles and different job specifications.
As work becomes increasingly automated, detailed skills requirements now form the basis of recruitment, rather than criteria for education and experience. Meanwhile, new occupations are appearing, requiring different, novel skillsets.
Recently, the rise of digital automation and Generative AI has led to signs of further disruptions, because these technologies can complement or even substitute the cognitive skills, previously seen as the preserve of human resources.
Since 2015, we observed there has been a five-fold increase in the demand for these skills...
In a new Citi GPS 'Skills that Pay' report, to which I contribute, two major changes were detailed that affect the labour market and were largely the result of AI development:
- the phenomenal rise in the demand for AI skills, and
- the rise of the collaborative leader as the 'glue-guy' of successful AI teams.
This came from a deep dive into the demand and supply of AI skills over the period 2015-2022 using global job-ad data along with profile-level skills of the existing workforce.
Since 2015, we observed there has been a five-fold increase in the demand for these skills as a percent of US jobs. While, globally, AI jobs grew nine times for Tech-AI jobs and 11.3 times for Broad-AI ones while IT sector jobs rose 3.7 times and the number of total jobs advertised grew 2.7 times.
Asia stands out as the region with the highest growth over this period, surpassing Northern Europe.
Since 2017, the industries most affected by these technologies have experienced spectacular increases in the demand for AI talent:
- manufacturing experienced a sevenfold increase in five years,
- management a fivefold increase, and
- warehousing a 21-fold increase.
At the other end of the spectrum, it is striking that healthcare AI-talent recruitment is five times smaller than the cross-industry average, generating a huge investment opportunity. New tech-hubs have been formed in the US, with the states of Texas and Virginia rising to the top spots for AI demand and surpassing New York and Washington.
Since 2017, the industries most affected by these technologies have experienced spectacular increases in the demand for AI talent
The supply of these hard-to-find professionals is also skewed across the US - with California host to almost a third.
The high demand for their skills, leads to candidates with next to no experience being hired for senior roles. More than half of Tech-AI professionals have less than six years’ experience, compared to an average of 10 years for all IT jobs.
In some states, there are more than 10 job ads per AI professional. The wage implications of this are important - and it makes the quest for talent a global endeavour, which usually has to incorporate remote and hybrid arrangements.
Casting a wide net, increases the pool of candidates and significantly reduces cost, making it a viable and often necessary option.
Skills shortages have also seen strong support for the importance of collaborative leadership in building a successful AI team. People who can coach, influence, negotiate, take strategic decisions and effectively lead the scarcity of AI talent, are in high demand.
The high demand for their skills, leads to candidates with next to no experience being hired for senior roles
As more tasks are gradually automated, the intersection of technology understanding along with the soft skills, necessary to make them thrive, are very thin on the ground. There is evidence of a wage premia for such leaders across all skills categories.
AI skills have been at the forefront of labour demand so far, especially as they are constantly shifting the technological frontier further and when combined with the hard-to-teach soft-skills that very few 'glue-guys' can deliver to their teams.
So, what can we learn from all this? Essentially, those looking to study IT or working as an IT professional may find brushing up on AI-related skills is probably the way to go if they are looking to specialise further. However, it is also important not to forget to build the soft skills needed to succeed in such areas.
We split AI-focused job-ad data into two groups: the Tech-AI jobs that relate to the technological skills that are necessary to “run, train and test” AI models; and the Broad-AI jobs that require an understanding of AI technologies but do not need the “hard” tech-skills found in Tech-AI.
We examined key skills, including Artificial Intelligence, machine learning, natural language processing and computer vision, along with tens of other AI specific skills.
This opinion piece reflects the views of the author, and does not necessarily reflect the position of the Oxford Martin School or the University of Oxford. Any errors or omissions are those of the author.