**5. The concept of "fusion" skills**

Research and discourse about the impact of the 4IR has, to a large extent as we saw earlier, focused on the aspect of substitution and automation: what tasks and activities smart machines currently are or soon will be able to perform and what the implications for the labour market are [17, 21, 56]. An alternative perspective has however been present by Daugherty and Wilson [5] in their book *Human + Machine: Reimagining Work in the Age of AI*. They argue that the above debate has been constructed around a separate focus on either tasks that are performed by humans or alternatively tasks performed by machines. As a consequence, an important range of activities is lost out of sight: hybrid activities where humans and machines closely collaborate – as exemplified in the case of robotic surgery. This is a radically different way of identifying not only the 4IR's or Society and Industry 5.0's skill needs compared with the production of lists of digital skills, but also the implication of these skill needs for engineering, as we explain below.

Employing a forecasting methodology, in common with the advocates of the substitution perspective, Daugherty and Wilson [5] nonetheless adopt a very different approach. Instead of asking the question – how might AI impact on jobs? – they ask – how might result in new jobs or new roles? To do so, Daugherty and Wilson [5] distinguish between three types of work activity: human-only activity, such as leading, empathising, creating and judging; machine-only activity, such as transacting, iterating, predicting and adapting; and human and machine hybrid activities. They sub-divide the latter into two categories: activities where humans complement machines, such as training, explaining, sustaining; and activities where AI gives humans "superpowers", such as amplifying, interacting and embodying. Based on this distinction about different types of human + machine hybrid activities, Daugherty and Wilson make the following inter-connected argument. Firstly, that: "the novel jobs that grow from the human-machine partnerships are happening in what they "call the missing middle – new ways of working that are largely missing from today's economic research and reporting of jobs." Secondly, the emerging human machine hybrid activities will require "fusion skills". Thirdly, the most important fusion skill will be to "reimagine" how AI can be used as a resource to transform working, living and learning. As conceived by Daugherty and Wilson [5], each of the skills they identify draws on a fusion of human and machine talents within a business process to create better outcomes. Their eight fusion skills are:


These skills are, unlike the digital skill list we presented earlier that merely constituted a series of additions to extant interpersonal and technical skill such as, data analytics, based on forecasts about how humans will in future work with machines. Daugherty and Wilson formulated their fusion skills by analysing extant human-machine interaction and identifying human-only and machineonly skills, and then identifying on the basis of the future deployment or development of AI the new kinds of interactions that could occur between humans and machines in the context of work. This approach is therefore also radically different from Frey and Osborne [17] and Muro *et al*. [21] who operated with a classic social science binary assumption – automation or continuation – of work. Furthermore, unlike the advocates of the substitution perspective who steer clear of discussing the implications of their forecasts for organisational strategy, Daugherty and Wilson ([5], p. 3) argue that in order for companies to gain the most value from AI they will need to "reimagine" their operations and identify the requisite fusion skills.
