A new report carried out by research agency Ipsos Mori into the current state of the UK’s AI labor market found that close to 110,500 job opening were posted in the past year for roles related to AI and data science. That’s more than double the number of vacancies registered in 2014, and a 16% increase from 2019, marking the highest year to date for AI jobs posted on the market. Every month for the past three years, between 8,000 and 10,000 roles were posted online, ranging from data analysts and software developers to research and development or even university positions such as lecturers and professors in AI and data science. Sectors where demand is the highest, found the report, are the education and financial industries. SEE: Building the bionic brain (free PDF) (TechRepublic) In other words, for prospective candidates, there is plenty to pick from. And more is coming: two-thirds of firms, found Ipsos, expect the demand for AI skills in their organization to increase in the next 12 months. The benefits are good, too. With a mean advertised salary of £54,800 ($77,388), jobs in AI and data science offer a wage premium of 22% compared to IT roles overall. But as alluring as the role description might sound, found the report, employers are struggling to fill their job openings. The talent pool for AI, it would seem, is not sufficient to meet businesses’ demand, and 69% of firms reported that they had found it difficult to fill at least one vacancy in the past two years. Much of the problem boils down to a lack of appropriate skills among applicants. More than two-thirds of businesses said they struggled to find candidates with the right technical skills and knowledge, while a significant minority of others (40%) reported a lack of work experience, as well as gaps in industry knowledge. So, what exactly should candidates for AI and data-science roles have on their CVs to convince future employers? Technical skills, of course, are key: businesses said that they were in search of applicants who understand AI concepts and algorithms, know programming skills and languages, and are familiar with software and systems engineering. A number of employers, said Ipsos, stressed the importance of deep learning in specialist roles, and of the need for candidates to know how to go beyond “low-level” AI. “We need people coming through the university system to learn from first principles how to create deep learning, neural network systems, rather than relying on off-the-shelf systems that are available through the big US companies,” said one micro-business owner. For Roger Woods, dean of research at Queen’s University in Belfast, who co-authored the report, the solution lies in creating dedicated courses that will train students to meet business’s need for deep technical expertise from an early stage. “Whilst mathematics and further mathematics A-level course material is being modified to reflect the needs of AI, there is a strong case for a dedicated AI/machine-learning A-level course,” Woods tells ZDNet. “This will act to provide a greater number of talent coming from schools with some suitable expertise.” At the level of higher education, too, things seem to be moving along. Ipsos found that universities offered over 700 undergraduate courses in AI, robotics or data science last year, compared to only 122 in 2019. “This will act to produce qualified staff, but, of course, there will be a three-to-four-year lag before any immediate impact will be seen,” forecasts Woods.
This AI-powered resume builder increases your chances of landing your dream jobEverything you need to know to land a job in data scienceDon’t let a bad tech stack hurt employee retention, use these tips to improve worker IT satisfaction
In the short term, many companies are banking on current or prospective employees training themselves while in employment, to ensure they are up to date with the skills that are required of them. An overwhelming 96% of the business leaders surveyed said that their staff were typically good at self-learning; employees, for their part, mentioned reading research papers and undertaking personal projects as examples of initiatives they took to boost their skills. Short courses offered by universities and commercial providers are also increasingly used to improve on-the-job training. Google, for example, provides a “professional machine-learning engineer” certification for $200, which assesses the ability to frame, architect and develop machine-learning models. On Coursera, users can take a six-course certificate for AI engineering, which is recognized by IBM, and which teaches the fundamental concepts of machine learning and deep learning. SEE: Developer burnout and a global chip shortage: The IoT is facing a perfect storm This type of short-term training also has the potential to lure in talent from other fields. “A more immediate response is to create one-year masters conversion courses,” says Woods. This was done successfully for software in the past, allowing bright people with a background in more general science and even humanities, to experience an extensive one-year training in the area. Ipsos’s report identifies an opportunity for people with unrelated skills to re-train and build up their proficiency in AI. The research mentions the example of a participant with a background in the arts, whose interest in logic and mathematics led them to pursue a masters in quantitative social science, followed by a PhD in an AI area. Another respondent did an undergraduate in life sciences and a master’s degree in public health, before working as a statistician. After teaching themselves some data science skills as part of their job, they moved to a role focused on data consultancy.
AI and data science jobs are hot. Here’s what employers wantWhat is Agile software development? Everything you need to know about delivering better code, faster Developer burnout and a global chip shortage: The IoT is facing a perfect storm
In other words, the bridge between AI and other fields can be easily crossed by motivated candidates. Woods, however, is keen to stress that garnering a high-level understanding of AI, and learning the technical skills that are so sought after by employers, inevitably entails some scientific background of sorts. “This is probably influenced by my role of academic and chief scientist of an SME,” he says. “But drop your mathematics at your peril. Fundamentally, many of the opportunities are based on a strong understanding of mathematics.” But it’s not only about technical skills. Ipsos’s report also showed that employers are looking for staff who can demonstrate softer skills, such as communication, leadership or management – and even, in some cases, an understanding of the privacy and ethical issues surrounding AI. Crucially, about half of employers said that candidates seem to lack “commercial awareness”. It is becoming increasingly important, in effect, to understand AI in the context of business, and not only as a scientific project; and many candidates are now lacking the skills to have a “real-world” approach of the technology. SEE: What is a software developer? Everything you need to know about the programmer role and how it is changing Bledi Taska, chief economist at labor market data company Burning Glass, who did not participate in the research, explains that two new kinds of skills are becoming valuable to those working in AI: human skills like communication, teamwork and collaboration; and what he describes as “business enablers”, which consist of understanding the bigger business picture. In other words, applying for an AI or data science job now also means knowing how to engage with the larger industry problem that the employer is trying to solve with machine-learning models. “Everyone is focusing on the technical skills, and how make a model go from 90% accuracy to 93% accuracy for example. But when they are asked to explain this in simple words to a client or team member, they cannot,” says Taska. Developer, manager, salesperson: as data science become central to more firms’ business models, those driving AI projects will be asked to go beyond the lab to have an understanding of the skills that make up the foundations of industry. “What we need to focus on, is developing expertise, and systems that solve real business problems in a more efficient way than before,” says Taska. “What we need and what is currently missing from the market is deep expertise both in the technical skills, but also in the human and managerial skills,” he concludes. An observation worth keeping note of, for the next time you revisit your CV.
