Workers and Machines #6

An ask for readers: 

The Government recently opened its consultation on workplace monitoring technologies, which we've been told will still be going ahead despite recent government shakeups. This is part of Labour's manifesto commitment to Make Work Pay. The TUC will be working on our response over the summer.  

We’re asking unions and workers to get in touch with real life examples and insights on how surveillance tech is playing out in workplaces. Please contact Isabella to share.  

Adam and Isabella 

What we're up to

A fair AI economy for London: The London AI and Jobs Taskforce's report, Shaping how AI changes work in London, launched yesterday. The report sets out how London can and should act quickly to shape the AI transition to help employers adopt AI responsibly and create new opportunities for workers – and ensures unions have a seat at the table. The mayor has accepted the report's recommendations.

Education Education Education: Connected by Data and the TUC worked with nine education unions to test how to catalyse the movement’s capacity to understand, negotiate and advocate on AI. After a seven-month stint running this "action learning" programme, we’ve published our key lessons from education unions building power on AI, setting out how this project shifted the dial on union work on AI.

AI is a policy issue: Adam represented the TUC at the Business and Trade select committee inquiry on AI and the future of work. The panel touched on everything from algorithmic pay-setting in the platform economy to the deskilling of programmers and the need to reorient the UK economy towards workers and away from corporate short-termism. You can find the full recording here.

AI is an industrial relations issue: Isabella spoke with Personnel Today on how HR and AI might be colliding. She said: “AI is often being wedged into workplaces by executives, while the people who are expected to use it are excluded from decisions about what's being implemented and why.” Read it here.

AI is an opportunity for workers to leverage: We made a guest appearance on the Institute for Public Policy Research (IPPR) Substack, arguing that counter to common narratives workers have more power on AI than often assumed. Read it here: Why workers actually have significant leverage on AI right now.

Movement movements

Big win in for performers: Equity has won AI protections for television workers in its new bargaining agreement with Teledwyr Annibynnol Cymru (TAC) – the trade body for the independent TV production sector in Wales. The deal includes stronger rules on digital recreations of performers, increased transparency on use of AI, and recognises the principle that tech should support creative work. This is one of the first agreements in the UK to establish specific protections governing the use of AI in TV.

Civil servants are hoping to follow suit: PCS want the Cabinet Office to agree to its model agreement on AI, which asks for clear worker protections, training, and oversight. The proposed framework “aligns with the shared interests of both sides: to make sure AI improves work, maintains fairness, and avoids decisions that could adversely affect staff or citizens.” Read the full list of asks here.

Welsh workers want their fair share: TUC Cymru’s Congress passed two resolutions on AI. One calls for a shorter working week with no loss of pay for Welsh workers in response to productivity gains from AI and automation. The other called for TUC Cymru to continue their work on AI through policy engagement and collective bargaining.

Workers face some hurdles organising at Google DeepMind: Unions say Google "isn’t engaging in good faith" with union negotiations. Workers say Google attempted to shut down dialogue around the unionisation bid, and – after opting to negotiate through Acas’s conciliation service rather than accepting the union’s request for voluntary recognition – senior management failed to attend the initial meeting.

Over 100 NHS workers say no to Palantir: NHS Analysts Together has sent an open letter to leaders of NHS England calling for an end to its contract with Palantir. Backed by union groups including UNISON, United Tech and Allied Workers Union, and NHS England Unite Members, the letter asks for an ethical procurement framework and consultation with workers around data infrastructure.

British Gas is axing 500 call centre jobs: The company says customers prefer using chatbots for routine queries, but GMB says call centre workers are already over-worked and underpaid. This isn’t just an issue for those workers who lose their jobs. When the easy issues get handled by bots, workers spend their days dealing with angry callers and intractable problems. Working conditions get worse and pay stays the same, while corporate profits increase. See above point on the resolution adopted by TUC Cymru on a shorter working week for one example of how gains could be shared instead.

Bookmark this: Lessons on building union capacity on AI

We urgently need to boost the movement’s capacity to understand, negotiate and advocate on AI. But with diverse technologies rolling out across sectors, how best to do that?

In partnership with nine unions representing education workers, Connected by Data and the TUC worked across seven months on an "action learning" programme to test how to catalyse the movement’s capacity to respond to AI from a worker’s perspective.

The project shifted the dial on individual union officers' confidence on AI, union structures, strategies and practical outcomes for campaigning on AI. Read about it here.

We will be taking this forward in a new project supporting unions on AI-related workplace issues. But first, we want to hear from you: what you're seeing, what you need, and what would actually be useful. In return, you'll get early input into a resource shaped around real experience, plus the chance to learn and connect with others across unions facing the same issues. This might include: a 30-minute interview, attending a workshop, and/or testing and review.

If you are a negotiating lead, organiser or national officer and want to be involved, get in touch!

In brief: When surveillance becomes task automation

Until ChatGPT was unloaded on the world in late 2022, the predominant focus for workers and unions when it came to AI or digital technologies was algorithmic management – the use of data-driven and automated tools for setting targets, "robo-firing" platform workers or surveillance at work. Now a new dynamic is unfolding: the use of surveillance to collect data to train AI and potentially displace workers.

The monitoring of workers to extract information isn’t new. It was most famously embodied in Taylorism – a management theory named after Frederick Taylor. His time-and-motion studies were developed to scrutinise manufacturing workers as they went about their tasks. Management’s goal was to identify efficiencies, break up tasks into more measurable components, and control the work process. Metalworker and socialist Harry Braverman saw it as part of a process of deskilling: as workers lost the ability to plan and control their craft and instead performed a limited set of tasks directed by management.

The AI version of this looks like Meta’s plan to collect data from workers' screens to train AI agents – now on pause due to worker backlash. Or factory workers and domestic cleaners wearing cameras to collect data for training humanoid robots. Related is Uber’s use of driver-generated data for automated vehicle rollouts, as is happening in London this year. Though a different mechanism, it's effectively the same process of data collection as when Large Language Models (LLMs) are trained on data scooped from the internet. The data is made by creative workers, lawyers, coders and then used to automate generation of images, legal briefings or lines of code.

Sometimes, though, companies move too fast to automate and must rewind, as Ford Motor Company has recently. Finding that its AI-enabled quality control system on the production line was defective, they have recently rehired 300 experienced engineers. In a vindication of workers’ skills and expertise, an executive said: “We didn't pay as much attention as we should have to the experience of our most knowledgeable engineers." He went on to say that “we needed to ensure that [the AI tools] were trained by the most experienced individuals", indicating that once the systems had been trained the workers would be let go again.

Varying forms of automation – from the loom to AI agents – have always been a feature of work. The challenge for the labour movement is to shape the change, resist where necessary and claim a share of the gains.

Events, etc.

Instead of events this month, we are bringing you our summer reading recommendations.

We Are Not Machines by Sarah O'Connor: an exploration of how AI is changing the character of work from a columnist, reporter and associate editor at the Financial Times.

The "Messy Middle" by Molly Kinder on Substack: a look at why policymakers can't afford to overlook "the long, hard stretch of job disruption between today’s mostly intact labor market and the post-AGI world of abundance that Silicon Valley keeps promising".