Employers’ Digital Practices at Work Survey: First Findings
Mark Stuart, Danat Valizade, Felix Schulz, Brendan Burchell, Richard Dickens and Jacqueline O'Reilly (2023)
There are significant gaps in our understanding of how employers in the United Kingdom adopt and use artificial intelligence powered tools and software.
This report provides a comprehensive analysis of a nationally representative survey of business establishments in the United Kingdom conducted between November 2021 and June 2022.
The survey offers both positive and concerning findings. Employer investment in digital technologies appeared to be positively associated with employment growth, future investment and employee involvement practices. However, only a minority of employers were ‘digital adopters’ with seemingly little appetite for future investment in AI. Employers’ investment in skills and training also appeared to be low.
If the UK economy is to shift its growth model towards some form of digital transformation, then addressing the emerging divide between digital adopters and non-adopters should be an urgent priority.
Investment in artificial intelligence and machine learning-enabled technologies
The use of traditional information and communication technologies was widespread, with desktop computers, portable devices and smart phones ubiquitous. More than 90 per cent of employers reported that they used desktop computers. In contrast, more novel forms of ICT hardware, such as wearables, were much less common, adopted by around a fifth (22%) of employers.
Just over a third of employers had invested in AI-enabled technology in the past five years. We label these ‘digital adopters’. Cloud computing and the internet of things (IoT) were by far the most popular types of new digital technologies (79% and 59% of digital adopters invested in these technologies respectively). Investment in AI-enabled equipment and apps was much less frequent.
Digital adopters were more advanced than digital non-adopters in the use of traditional ICT technologies, especially wearable devices and customer relations software.
Digital adoption was uneven. Firm size was significant; one in two firms with more than 100 employees were digital adopters compared to around one in three among small enterprises (with less than 50 employees).
Industrial sector also mattered. For instance, around two thirds of employers in Public Administration and Information and Communication had invested in digital technologies, while only 22 per cent of those in Accommodation and Food Service Activities and 30 per cent in Education had invested.
Improving efficiency, productivity and product and service quality were the main reasons for investment. Digital technologies were mostly applied in the organisation of production or service delivery, and quality control and monitoring. Non- adopters typically identified their area of business activity, wider business risks and the nature of skills demanded as the key reasons for non-investment.
Digital adopters appeared to be very open to future investment. Six in ten reported that they would invest further in digital technologies in the next two years. In stark contrast, just one in ten digital ‘non-adopters’ intended to invest in the next couple of years.
COVID-19 appears to have accelerated digital investment. Seven in ten digital adopters cited the pandemic as a driver of increased investment in AI-enabled technology.
These findings indicate a tangible risk of a growing digital divide in the UK economy where a significant proportion of UK employers could be left behind the digital transformation.
Employers’ use of data analytics
A little less than 40 per cent of employers used data analytics. Larger firms were more likely to use data analytics. Use of data analytics was more prominent amongst Financial and Insurance employers, Energy suppliers, and Information and Communication employers.
Almost eight in ten organisations using data analytics reported increasing reliance on data in the past five years. Two thirds of digital adopters expect this trend to continue. Around half reported that the use of data analytics had increased since the COVID-19 pandemic.
Importantly, six in ten data analytics users indicated that employee records and job applications were key data sources. Sales and performance records, transactions and customer feedback were the most widespread data sources.
Data analytics were most commonly used for marketing, finance, the organisation of production and customer relations. In terms of human resources, over half of analytics users applied data analytics in performance assessment and slightly less than half used data analytics to allocate work and optimise working time. Just over a third of businesses deployed data analytics in recruitment and selection.
Cases of semi or fully autonomous algorithmic decision-making were rare. Between one-fifth and a quarter of employers using data analytics employed algorithms as decision-making tools in strategic management and human resources.
The main reasons for employers not having used data analytics included: scepticism in the capacity of algorithms to solve real-life business problems; lack of skills; and the lack of a business case for using data analytics.
Staffing and human resource management practices
Employers had an overall positive employment outlook. One in four reported increased staffing levels over the past five years, and one in six expected to grow in the near future. Adoption of digital technology was positively associated with reported increases in employment. There was no association between digital adoption and changes in working hours, past or future.
It was rare for employers to have trained all their staff in the previous 12 months. Only six per cent of employers had put all of their employees through some formal digital skills training, although on-the-job training was reported by nearly a quarter of employers.
Digital adopters reported higher incidence of digital skills training. The widest gap between digital adopters and non-adopters was in relation to informal, on-the-job-training. Around a third of adopters, compared to one fifth of non-adopters, relied on this type of training for the acquisition of digital skills.
Digital adopters reported mixed results in relation to employee autonomy and discretion. Adoption of AI-enabled technology was associated with greater employee control over their working time and higher incidence of teamwork.
Digital adopters more frequently used algorithmic control over work organisation (three times more than non-adopters). However, it should be noted that such an approach was relatively uncommon: just six per cent of digital adopters reported the use of algorithmic control. Digital adopters were more likely to use machines and algorithms to determine the pace of work.
Almost every third employer reported the use of customised software apps to monitor working hours and absenteeism. Such apps were not associated with working time autonomy but showed a positive correlation with machine control over the pace of work and algorithmic control over work organisation.
Employee autonomy or control was largely linked to job type. Employee control over working time was prevalent among managerial and professional occupations. Algorithmic control was more widespread in routine occupations.
Digital adopters were more likely to have negotiated or consulted with their employees across all key working practices. This includes working time, pay, recruitment and selection, and the use of data, including data on employee performance.
Employers’ decision-making structures were most likely to involve employees in relation to training and skills development. Employee voice was least apparent around investment decisions related to new technology.