Berkers notes that Schimmelpenninck acknowledges the technology is man-made, yet at the same time adopts a deterministic discourse that she describes as a sales pitch by Big Tech. The fact that AI is developed and deployed by people means there is room for responsible choices. Companies like Anthropic and OpenAI develop the technology, but they do not control the narrative. According to Berkers, this line of thinking does little justice to the many organisations, professionals, researchers and policymakers working on responsible AI applications.

To illustrate her argument, Berkers points to concrete examples. Research involving customer service employees who worked with AI support showed that they became more productive, learned faster and found greater satisfaction in their work. At Dutch healthcare organisation ZorgDomein, a chatbot named Zoë took over dull and tedious tasks from staff, resulting in greater efficiency and higher job satisfaction. In practice, work did not disappear as a result but became more challenging. Berkers also underlines that AI is a broad concept: alongside large language models such as ChatGPT, machine learning is also used for fraud detection and for the smarter deployment of MRI scanners.

At the same time, Berkers acknowledges the genuine risks. The environmental impact of AI, indications that it undermines critical thinking, the intensification of work and its misuse by organisations as a pretext for redundancies are all serious concerns. Yet here too she adds nuance: organisations that deploy AI to replace people often come up short. Klarna had to reverse the dismissal of seven hundred customer service employees because the AI failed to deliver on its promises. The same technology, similar work, but a completely different outcome compared to ZorgDomein. The difference lies not in the technology itself, nor in Big Tech, but in a mentality.

In the context of healthcare, Berkers sees the deployment of AI as unavoidable. Due to staff shortages and an ageing population, workloads are set to rise sharply in the coming years, with potentially disastrous consequences if nothing changes. Without AI, part of that work will simply no longer be manageable. She references French economist Thomas Piketty, who questions whether some of that work might not need to be done at all anymore. This calls for a human-centred and responsible approach to AI design and deployment, a view shared among researchers in the field of work and AI.

Berkers concludes by reframing the question Schimmelpenninck poses. Not: why is AI needed, but: what is AI needed for? Sometimes AI solves something, often it does not. Making deliberate choices is therefore central to her message, as individuals, but also within teams, organisations, professions and as a society. AI is not a force of nature, and its implementation is a choice that people make themselves.