Gorodenkoff - stock.adobe.com

Welsh workers’ AI experiences shaped by uneven power dynamics

Workers in Wales are being negatively affected by the unregulated introduction of artificial intelligence into workplaces due to asymmetric power dynamics

Asymmetric power dynamics at work are fuelling Welsh workers’ negative experience of artificial intelligence (AI), says Trade Unions Congress (TUC), and make it difficult to challenge the imposition of new technologies in the workplace.

In a review of how workers and trade unionists across the country are experiencing the technology, Wales TUC found that their efforts to understand and challenge AI in the workplace are being hobbled by a lack of access to useful information on the systems and a near-total exclusion from decision-making processes around its deployment.

For example, it said that while employers themselves may not have a full understanding of the functions and effects of their AI tools, the lack of information available to workers – both on AI in general and in relation to specific systems deployed by their business – is undermining effective negotiations.  

Wales TUC said workers have had limited success when it comes to shaping the introduction of AI, as these processes are also characterised by an absence of participation and transparency, and there is no legal obligation on employers to consult their staff about it.

“Reps cited having limited purchase on shaping the introduction and deployment of technologies, and difficulty in breaking through to managers and employers about the human decisions behind seemingly ‘objective’ algorithmic targets,” it said.

“According to reps, technologies were being introduced into the workplace by employers with the stated goal of improving accuracy and efficiency. However, a weak understanding and improper application of the technologies, and failure to incorporate worker insight by employers, was leading to an environment where workers are further disempowered.”

Feeding into this dynamic, it also documented how Welsh workers are being negatively affected by AI-powered workplace surveillance and automated decision-making, because employers are generally more likely to trust the technology than the people it is monitoring.

For example, workers subject to electronic surveillance reported it was having a “dehumanising effect” on them due to the ways in which disciplinary measures were being taken of the back of data-driven systems that fail to take into account the full context of a situation.  

This, in turn, creates situations where workers are being forced to justify themselves when pulled up by management about the system’s outputs. “The risk is that we are losing the human elements in HR through automation and remoteness. There is no benefit of the doubt, no grey area,” said one worker at a large manufacturing site.

Wales TUC added that while workers’ disempowerment on AI was less pronounced in sectors such as manufacturing, where there has been a history of collective organising around technology issues, workers’ ability to exercise their voices and meaningfully participate in tech deployment decisions is still generally low across the board.

“AI is not one technology. It affects every workplace differently. Our findings are deeply troubling and demonstrate that conversations about AI and work are not academic parlour games – the impacts are real, often negative and are happening now”
Shavanah Taj, Wales TUC

“AI presents novel technical, legal and operational challenges that threaten to deepen power asymmetries in the workplace and wider economy,” it said. “However, this dynamic should be seen in a general context of some of the harshest laws governing industrial relations in western Europe, and employment rights that are not designed to empower workers to be active stakeholders in their workplaces, regarding AI or any other issue.”

It said while workers from sectors outside manufacturing are further behind in developing their technical expertise and negotiating tactics to deal with these issues, work is being undertaken by unions to better understand how they can successfully collectively bargain on AI given its variability and relative complexity.

Off the back of its review, Wales TUC is calling on the Welsh government to use its devolved public sector to lead by example on worker inclusion, which it can do by updating existing Workforce Partnership Council (WPC) agreements to make worker engagement a reality.

It is also calling on the Welsh government to develop, promote and monitor best practice guidance for worker participation on tech deployments; and use what influence it has over non-devolved processes to shape AI-related legislation – such as the Data Protection and Digital Information (DPDI) Bill – to the benefit of workers.

In April 2023, TUC assistant general secretary Kate Bell said despite the risks that AI could lead to widespread workplace discrimination, the UK government is refusing to put in place the necessary “guardrails” to safeguard workers’ rights.

“Instead of clear and enforceable protections, ministers have issued a series of vague and flimsy commitments that are not worth the paper they are written on,” she said in reference to the AI whitepaper setting out the government’s approach to regulation.

“They have failed to provide regulators with the resources they need to do their jobs properly. It’s essential that employment law keeps pace with the AI revolution. But last month’s dismal AI whitepaper spectacularly failed to do that.”

Regarding the DPDI, Bell said the government was “watering down important protections… [which] will leave workers more vulnerable to unscrupulous employers”.

For trade unions, including Wales TUC itself, the review added there should be a concerted effort to produce targeted resources and guidance to help workers through the process of challenging AI deployments at work.

“In the absence of clear requirements for employers to divulge what technologies have been deployed, the Wales TUC will [also] need to maintain information-gathering capacity, to inform opportunities for action,” it added.

In turn, it also highlighted the need to develop new practices and policy initiatives to empower workers in the face of AI.

“AI is not one technology. It affects every workplace differently. Our findings are deeply troubling and demonstrate that conversations about AI and work are not academic parlour games – the impacts are real, often negative and are happening now,” said Wales TUC general secretary Shavanah Taj.

“We will fight to ensure that our workers have voice and control over any changes to their workplaces and working conditions.”

Read more about artificial intelligence at work

Read more on Business intelligence and analytics