Employees should be empowered to refuse the use of artificial intelligence (AI) tools at work, says the tech worker union, after finding employers are using the technology to intensify workloads, eliminate junior hiring and rationalise redundancies.
Conducted by the United Tech and Allied Workers (UTAW) union – a branch of the Communication Workers Union (CWU) – the AI workers’ inquiry 2026 explores how AI tools are reshaping labour practices from the perspective of workers throughout the tech sector.
Focused less on the technology’s functionality and more on who controls and benefits from it, the inquiry found that AI technology was “a largely negative force” in the workplace, leading to a widespread deskilling and degrading of expertise, expanded surveillance, less accountability for managers and weaker job security.
The worker-led inquiry also found that AI, despite being touted as a productivity-boosting technology that can reduce working time, is being used in practice to increase and intensify people’s workloads.
All of this is taking place, it added, without workers having a meaningful say over how AI technologies are introduced.
“Workers stressed that positive uses depend on their ability to decide for themselves how to engage with AI,” said a report on the inquiry’s findings. “The central problem is not simply whether AI works well, but how it is introduced, controlled, measured and used by employers. Put simply, it is useful only when one gets to choose if and how to use them.”
It added that workers must be actively empowered to defend and extend their rights over how AI is used, which should include being given “enforceable power” over how it is deployed in the workplace, as an antidote to its unilateral imposition from management.
‘Testing ground’
UTAW highlighted how, according to inquiry participants, senior management teams do not understand the tools they are mandating, and are using staff as a “testing ground” for tools that often either slow down or worsen the work they claim to speed up or improve.
The unilateral imposition of AI, UTAW added, is also part of a wider trend where AI is being used to reduce headcounts and avoid hiring, while distributing the remaining work to fewer employees under tighter deadlines.
“At first glance, companies are carrying out redundancies (both mass layoffs and small regular redundancy exercises) and claiming this is because ‘AI is replacing the need for workers’,” it said.
“Workers on the inside report that layoffs are more often to maintain revenue in the face of rising inflation, to increase company valuation by manipulating financial indices, or even to cover the cost of GenAI [generative AI] platform subscriptions.
“These so-called AI efficiencies are an excuse to obfuscate a reality where skilled staff are replaced with new tools that do not work well while workers who survive redundancy pick up the ever-intensifying slack.”
For those at the start of their careers, UTAW also highlighted how AI is undermining their ability to develop skills, because they are either becoming too reliant on the tech themselves, or their employer is outsourcing their work – which would help them develop foundational skills – to automation.
“Junior workers and those without strong foundational skills are becoming overly reliant on tools such as agents and assistants,” it said. “There is no incentive or time for juniors to learn the necessary technical and critical thinking skills; work intensification means they are increasingly being denied a sounding board to chat through problems.
“Employers are using LLMs [large language models] as a substitute for training their employees, obscuring gaps in subject-matter expertise and skills … The long-term trajectory is a loss of expertise and skills across the sector.”
Policy proposals
As a result of its findings, UTAW has made a number of policy proposals for AI at work.
These include protecting employees from disciplinary and performance processes which are based on how much they are using (or not using) AI tools; strengthening redundancy protections so employers cannot use AI adoption as cover for offshoring or outsourcing; and introducing a statutory duty on employers to “meaningfully consult” workers before AI tools are deployed.
It added that, where collective bargaining agreements exist, employers must have a duty to negotiate on AI deployments and their associated effects.
Where these agreements do not exist, UTAW said that workers must still be given a right to refuse AI “where there are professional, ethical, environmental, equality, accessibility, safety, religious, mental health or quality concerns” around its use.
The codification of a right to refuse should also be accompanied by protections for workers that refuse, question or otherwise criticise an employer’s use of AI.
A number of other proposals also called for increased transparency around AI deployments at work.
This includes more consistent disclosures on its environmental impacts, such as “energy use, water use, datacentre dependency, infrastructure demand and supply-chain impacts”; the publication of various pre-deployment impact assessments; and prohibiting the use of opaque, AI-driven work evaluations.
To ensure employers are held accountable, UTAW concluded there is a need to establish a new regulatory authority on AI at work, which has statutory powers to investigate, enforce and sanction both employers and their technology providers.
“The authority should have the power to require employers to disclose AI systems, pause deployment, carry out impact assessments, consult workers and their unions, remedy harm, and withdraw deployed systems,” it said, adding that any governance board “should include expert worker and trade union representation equal to the number of representatives of employers, civil servants and academics”.
