Recensione a S. Jeffares, The Virtual Public Servant. Artificial Intelligence and Frontline Work

Control, cost, convenience and connection are the four key policy problems related to front-line public administration work that AI might help to solve. Whilst technology has traditionally helped public management to control processes, staff and costs, while fostering efficiency and meet consumerist expectations by the public, AI solutions promise to overcome the problem of connection, “developing agents that can sense emotions, respond appropriately and continue to learn and adapt”. The book by Stephen Jeffares – a prominent researcher in the domain and a Senior Lecturer of the School of Government at the University of Birmingham – digs deep into this matter by providing an analysis of how AI solutions are massively gaining momentum in the daily life of public offices around the world, and how one should approach their design.


The ultimate goal of the Author’s research is to understand how policy actors play a role in “articulating support or opposition to the advancement of AI and related technology at the frontline of public service”. The study leverages on an extensive number of interviews with public servants and managers, online reviews by people who had been in direct contact with a public agency either physically or via other means, as well as on marketing materials used by the industry and a database listing AI solutions developed for public authorities. This rich material is then organized and analysed according to the Q methodology, to systematize the subjective perspectives provided by research participants.

The study remarks that face-to-face encounters in the public domain have become increasingly rare. A number of factors have been contributing to this outcome, namely scarce human resources and shrinking budgets, complexity of needs to be addressed and increased customer expectations. Technology is largely seen as a complement to human agents and interactions, enabling also a fostered managerial oversight on how the administration is capable to deal with citizens’ requests: a sequence of remote contact means are used, ranging from call centres to trained bots providing a set of template answers to a certain range of potential interactive scenarios.

Nonetheless, while remote contact enables control, efficiency and consistency, contact centres remain among the most complained about forms of communication – especially when public agencies outsource activities to external contractors. In this respect, aside from speeding up operations and increasing efficiency, AI solutions support remote operators by providing guidance on how to deal with repetitive situations and demands by citizens, ultimately enabling the provision of more personalised services meeting the latter needs.

In addition, public administrations increasingly rely on AI solutions aimed at incentivising citizens’ self-care and self-service, develop the use of social media as part of “corporate communication strategies”, and harvest data enabling to interpret behaviours, understand community needs and allow increased personalisation. Jeffares provides several examples of bots and agents deployed on the frontline of public services, showing how the human and the digital may interact and work together or in place of one another.

In this respect, one of the key suggestions emerging in the last chapter of the book is the need to engage public services staff in the design of AI solutions to be applied in frontline work. Given that the use of AI in public services is already a fact, and the future will certainly bring more applications, the Author stresses the need to engage staff in “meaningful conversations about the future of work, the possibilities, the opportunities, skill and knowledge requirements. It requires leaders to resist the temptation to become part of the tech industry’s marketing machine, to get nominated for the career-enhancing industry award”.

The research has also helped to identify three viewpoints on the use of virtual servants.

The first one contends the need to protect face-to-face services, stressing that people “want care and compassion and challenges from digital technologies, longer life expectancy and loneliness add up to making face-to-face work ever more important and it should remain possible in their service”. This perspective recognizes that social contacts can make a tremendous difference to people’s lives, that technology can be useful but does not match for face to face, and that technology would not permit passive monitoring, or virtual agents with deeper understanding or fairer than human workers.

The second viewpoint is in contrast to the former, and it is mostly provided by interviewees in managerial positions or involved in the development of new technologies. It emphasises concerns for the expected continuous access by the public to services, coupled with a heightened awareness of the need to deliver value for money for taxpayers. Technology is seen as an enabling factor to reduce the cost of service delivery, shrink waiting times and help automate the more simple parts of the service. AI would help also to secure consistency in the service and reduce incidents of poor service.

The third perspective sits somewhat in the middle, sharing a certain enthusiasm for technology whilst underlining that the latter would ultimately help “empower people and offer them more choice in how, where and when they access a service”. Technology is an enabling factor in the sense that it empowers citizens “to serve themselves when they can, to work with AI and related technology when it most appropriate, to foster independence and filter out unnecessary contact”, thus ultimately reducing “a culture of dependency”.

The challenge remains to listen and accommodate all these perspectives, ensuring that they all help to design the future of public services.

Licenza Creative Commons
Quest’opera è distribuita con Licenza Creative Commons Attribuzione 4.0 Internazionale