When we began our studies in August 2019 there was a lot of media hype about robots taking jobs following studies claiming almost half of American workers could find themselves surplus to requirements as digital technologies including AI, robotics, etc., were getting smarter at doing our jobs. At the beginning of the study the team’s other concern was how to manage hundreds of face-to-face interviews across seven countries covering North America, Europe and Asia.

There was enough other research, including our own earlier investigations, to call into question the immanent prospect of technological unemployment on a mass scale. But we did not anticipate the Covid-19 pandemic which locked down the project team before international interviewing got started. We were saved by Zoom and related advances in communication software which meant we were able to move online, along with the rest of the world. Given our focus on the future of work, the race to remote working and learning was a global natural experiment in doing things differently facilitated by new technologies.

Without the pandemic there is no way we’d have the hybrid models that now characterise the working lives of millions of people. And what is less visible is the shift from labour arbitrage to task arbitrage within global value chains where nothing moves apart from online parcels of work delivered virtually by people you’ll never meet. It also meant social research could be done differently. Our online interviews were surprisingly good, although I remain a strong believer in face-to-face interviews within context, as you need to get a sense of a place, along with those informal chats, especially if in settings which are not familiar, so hybrid models of social research are likely to be the future.

As a result of the lockdown, we extended our studies from three to four years to give time to finish extensive fieldwork even with the use of Zoom. At three years in, we were beginning to formulate our key finding, which in many ways were consistent with our earlier research as we proposed a ‘job scarcity’ rather than ‘labour scarcity’ lens for exploring the future of work, because we were not convinced the fourth industrial revolution would create more and better jobs than it destroyed. We could see how companies, driven by cost efficiency, were using digital platforms and talent management strategies to cheapen the price of knowledge, in part by trying to capture various aspects of it in software, creating a relatively small cadre of talent with permission to think, distinct from the rest of the workforce.

At that point in our deliberations, what we were studying seemed more ‘incremental’ than ‘transformative’. This was also true for our analysis of AI and big data tools we included in our analysis. Could they identify emerging business trends and how could they inform our understanding of AI policies on a global scale without having to read hundreds of documents? The results were disappointing as the technology was good at identifying where to look but of little use helping us understand what we were looking at. That was until November 2022 when OpenAI launched ChatGPT.  I wonder if we’ll soon be making a distinction between BC (before ChatGPT)/ AC (after ChatGPT)?

The technological possibilities of generative AI had been known for some time and many other companies were working on similar large language models (LLMs). But it was the user-friendly interface of ChatGPT that caught the public imagination and for us the realisation of its profound implication for the future of learning, research, and work. The creation of content in text, image, video, audio, code, etc. is what makes today’s digital innovation transformational. On a recent visit to Silicon Valley to interview those at the forefront of these developments, there was little doubt in the Valley that generative AI is the next frontier in the future of business and work. As one Valley veteran told us: “we’re coming after every industry”. But we were equally perturbed by inequalities we witnessed in San Francisco and the way everything was viewed by as engineering problem, with little sense of the social or ethical implications of these new power tools. The strap line for San Francisco lining some of its streets reads ‘ahead of the curve’, but if it is the future it needs to come with a serious health warning.

These issues where not lost on the corporate leaders we’ve interviewed for this research programme, especially in the last six months. But the productivity potential of generative AI often overshadowed its social consequences from a business perspective, which was viewed as beyond their control, and ‘something for government.’ But the reality is it’s something for everyone as we have spent the last twenty years preparing for the wrong industrial revolution.

Our Final Report will be launched at the Digital Futures of Work Global Conference in Singapore on 1^st November 2023 and will be available at https://digitalfuturesofwork.com/