We Are Not Machines

Louise Marston reflects on Sarah O'Connor's new book on the impact of automation and AI on work - and how it speaks to the workplace challenges that the WorkerTech Fund is trying to address

Last week, Sarah O’Connor came to the Resolution Foundation to launch her new book ‘We Are Not Machines‘. Fans of her FT column, and her AI newsletter, The AI Shift, co-written with John Burn-Murdoch, will not be surprised that the book is both insightful and deeply human. Through the stories of workers already being affected by automation, Sarah describes the choices that we all make when faced with changes to our jobs.

The book brilliantly shines a light on something that we’ve also started to understand through our WorkerTech programme: that it really matters how you use and implement technology at work. That the difference between it helping and hurting workers, is the care you bring, the questions you ask, and the way you go about it. We’ve been supporting WorkerTech companies for over five years now, through the Resolution Foundation and now also through the WorkerTech Fund.

There were three lessons that I understood from Sarah’s book about how we can make automation successful and pro-worker – lessons that I also recognise from the WorkerTech companies that we have supported:

1. Understand the users

Sarah told us about Swedish mine workers who were able to negotiate the introduction of automation to their workplace. By understanding the jobs that people do and what they value, what their concerns are, and how they work, the technology can be introduced and boundaried in a way that works for everyone.

In our portfolio, Earlybird’s software is designed for workers delivering employment support, i.e. those helping others to get into work. Earlybird’s Co-founder and Chief Technology Officer, Boris Bambo, calls their approach ‘going deep’. He says: “Going deep is a choice but most people don’t make it. The difference between a good product and a great one lives in the depth.”

Earlybird has designed software that works for the humans, not the other way around. Their software helps to gather information, and transcribe meetings, cutting down admin time. A participant with Seetec employment support, one of Earlybird’s clients, says “The reason my adviser and I have formed such a strong relationship in just two months is that she’s 100 percent present in our sessions. She doesn’t type notes or fill out forms on her laptop, which means we maintain eye contact. This makes me feel truly supported and heard.”

2. Keep the human

We can all appreciate the lower costs and better access that automation can give us. Who would now choose to go back now to a world without Google Translate? But when we make things easier for users, we sometimes make it harder for workers: Sarah’s book makes clear that there’s a real human cost to how we choose to use these tools. Some human translators of film and TV work are finding their jobs made more difficult, less satisfying and paying less. Choosing how the work is divided between tools and humans is important.

At Valla, the AI software makes it quicker and easier to evaluate a potential employment grievance, but their service also provides community and empathy, and supports legal coaches to have better conversations, speeding up the preparation and post-call work. One of their clients, who won his case at employment tribunal, says: “They didn’t just provide legal knowledge — they created a community. I connected with others who had been through similar experiences, making me feel less alone.”

3. We can (and should) fight back

Sarah ends the book on an optimistic note: the future of work can be better - “but not without a fight”. People, policymakers, unions, progressive employers all have a role to play in making that future. We’ve worked with Organise for years as they have created a platform for workers to find and support each other to make change in the workplace.

Last year, Organise hosted a campaign organised by the drivers at Tesco’s Dagenham depot, who were being asked to have AI-enabled CCTV installed inside their cabs (in addition to the existing 4D CCTV cameras on the vehicle exterior). In striking contrast to the mineworkers experience that Sarah describes in the book, they didn’t feel engaged or consulted in this process. Their Organise petition says: “We urge Tesco Dagenham to reconsider the decision to install AI-powered CCTV cameras that continuously film and monitor drivers in the trucks. Instead, we propose that alternative measures be explored that ensure safety without compromising our privacy and trust.” They won this campaign and forced a re-think.

This year, Organise have been developing more tools to help workers to identify changes caused by AI, and they are supporting individuals to work out their next steps – to do the fighting back that is needed in big and small ways to make that future happen.

Far too much of the AI debate is centred around an inevitable job-apocalypse. But that grim future is by no means certain. The examples in Sarah’s book, and in our WorkerTech portfolio, prove that we can achieve better outcomes for workers. And we all have a role in making them as widespread as possible.

Louise Marston is Director of Ventures at the Resolution Foundation – you can follow her on Bluesky here.