We use cookies to personalise and enhance your experience on this site. By clicking Accept or continuing to use the site, you agree to our use of marketing and social media cookies as described in our Privacy Policy.
Accept and close
Are there winners and losers in automation? Or is technological change a tide really ‘raising all boats’?
New research from IFOW reveals that the benefits and risks of workplace technology are, in fact, unevenly distributed across demographic groups, occupational grades, firms, and sectors in systemic and significant ways. These experiences are strongly related to the different levels of awareness workers have about their rights at work.
These uneven automation experiences at work raise fundamental questions: what sort of institutions do we need to ensure that technological transformation is equal, and can be fairly anticipated, monitored, and managed by our institutions?
Automation is often framed as a simple story of job displacement. But workplace technologies actually produce a diverse range of outcomes.
Some technologies, such as high-discretion augmentation, can enhance workers’ autonomy, productivity, job quality, and bargaining power. Others, like low-discretion augmentation, can have the opposite effect, increasing surveillance, reducing workers' control, and limiting opportunities for progression.
The research we've recently published shows uneven patterns in the ways workers are exposed to these different forms of automation.
We found that workers from Asian backgrounds are more likely to experience automation that results in work intensification or displacement, compared to those of white ethnic background. Black, Caribbean and African workers report higher exposure to surveillance and ‘matching’ technologies at work, which can make the gigification of work possible, undermining fair pay, terms and conditions. Interestingly though, this group also reports higher exposure to technologies that enhance their autonomy and learning, suggesting high-discretion augmentation and greater technological control can coexist.
Existing power hierarchies at work can also determine the quality of automation experience.
Workers in higher-level roles are more likely to experience automation associated with greater autonomy, improved job quality and wage-enhancing opportunities than those in lower occupational groups.
These disparities indicate that, without institutional intervention, technological adoption risks reinforcing and deepening existing inequalities across the labour market.
Experiences of automation at work also depend upon workers’ differing perceptions of the rights and protections available to them. That’s why we looked at ‘perceived rights’: workers’ overall sense of protection and power at work. This is not the same as measuring their formal legal entitlements or knowledge of employment law. Rather, it provides a window into workers’ legal consciousness, their perceived ability and willingness to raise concerns, and trust in workplace institutions.
Here too, our research reveals important emerging inequalities. Managers, directors, and senior officials report the highest perceived rights, followed by skilled trades. At the other end of the labour market, workers in elementary occupations, process and machine operative roles, and caring, leisure, and other service occupations report the lowest levels.
This difference in perceived rights should therefore be monitored by governments as the technological transition unfolds, resolving disparities in access to justice and perceived fairness, which can prove critical for broader institutional trust and social cohesion.
Perhaps our most important finding is that workers’ perception of their rights and their automation experiences interact differently depending on the wider environment of worker representation and HR culture.
We know that not all automation is equal, and that workers with stronger awareness of their rights are more likely to experience positive forms of technological augmentation. Even where workers report no access to representation mechanisms, those who perceive high workplace rights are more likely to experience enhanced autonomy from technology and appear more protected against intensification and displacement.
But having a strong perception of workplace rights alone is not enough.
Trade unions play an important role in strengthening workers’ understanding of their rights and, crucially, providing the collective capacity to exercise them. Our research supports the importance of this role: access to unions is associated with stronger perceived rights, and union representation does not weaken the protective relationship between perceived rights and automation outcomes.
Other forms of workplace voice, such as employee forums or consultative committees, can also provide valuable channels for workers to influence organisational decisions. But they perform a different function and should not be treated as substitutes for collective representation. Our findings suggest that non-union representation mechanisms are not uniformly protective: at very low levels of perceived rights, they can be associated with poorer outcomes for worker autonomy, while at very high levels of perceived rights they can coincide with greater exposure to some harmful forms of automation.
The implication is therefore not a choice between unions and other forms of voice. Strong rights, collective representation and meaningful workplace participation can play complementary roles in shaping technological change. Workplace voice is most meaningful when workers have the institutional support and power to make that voice effective.
The wider HR context matters too. We find that employee-centred HR philosophies are strongly associated with good automation, particularly when employees perceive stronger rights. But it is not uniformly protective: at higher levels of perceived rights, employee-centred HR is also associated with greater exposure to harmful automation outcomes. This reinforces the need for employee-centred HR practices to translate into meaningful collective representation in technology decisions, rather than relying on engagement alone.
As a result of these uneven outcomes, governance needs to look ahead, not react once technologies are already embedded in workplaces.
This means building intelligence and intervention across the innovation ecosystem - from research and design, to development, procurement and deployment – to understand who benefits from technology and who bears its costs.
There are practical ways to do this.
Particularly in the UK context, the AI and the Future of Work Unit, and forthcoming Institute of AI and the Economy, should monitor automation archetypes across demographic groups, occupational grades and sectors, helping inform regulation, public investment and industrial strategy that supports positive technological outcomes as well as broader institutional trust and social cohesion.
Sandboxes could also test new governance frameworks in practice, while funding bodies such as UK Research and Innovation (UKRI) and Innovate UK could routinely assess the workplace impacts of the technologies they support, to ensure public funding favours forms of automation beneficial to workers.
More broadly, employers and HR practitioners also have an immediate role. HR should be involved alongside technology and operational teams in the procurement, implementation and ongoing review of workplace technologies, assessing their effects against Good Work principles. Employee-centred practice should translate into meaningful worker involvement in decisions about how technology affects autonomy, learning, workload and monitoring –not simply consultation tokenised to facilitate adoption. Existing worker voice channels, including trade unions where present, should be used to identify and address harmful impacts.
The overall message is clear: a more integrated approach to governing workplace technology is needed and worker voice must be given institutional backing.
A more integrated model of workplace governance should strengthen rights, collective representation and meaningful participation together. This includes accelerating access to trade unions and collective bargaining around the adoption of new technologies, as much as strengthening workers’ awareness of their rights, in order to deliver meaningful protection without waiting for legislation to catch up.
Alongside collective representation, Workplace Technology Forums and dedicated technology representatives could provide structured channels for ongoing reflexive dialogue between employers, workers and representatives throughout the design, procurement, deployment and review of workplace technologies.
We are honoured that this work has been published with the generous support of Innovation Resource Center for Human Resources (IRC4HR) and Alex Ferry Foundation - two organisations with shared commitments to fairness, dignity, and justice at work. Both organisations’ histories reflect the enduring importance of improving working lives through better institutions.
IRC4HR, a US-based foundation, traces its origins to a defining industrial conflict of the twentieth century. Following the 1913–1914 coal miners' strike in Colorado and tragic Ludlow Massacre, John D. Rockefeller Jr. established the organisation, now known as IRC4HR, to improve employment relations through better approaches to human resource management. As IRC4HR marks its centenary, that founding mission - to advance knowledge and practice in workplace relations - remains strikingly relevant as organisations navigate the opportunities and challenges of AI.
The Alex Ferry Foundation commemorates a Clydebank-born trade unionist who devoted his career to improving the lives of engineering and shipbuilding workers. As General Secretary of the Confederation of Shipbuilding and Engineering Unions, Alex Ferry was instrumental in securing the 37-hour working week. The Foundation itself is now supported by funds raised through that campaign for shorter working hours, creating a lasting legacy that continues to support research into the future of work.
Both histories carry a common lesson: technological and economic progress alone do not guarantee better work. Job quality, working conditions and fairness depend on institutions capable of shaping innovation in the public interest. That lesson is every bit as important in the age of AI.
