Background

All kinds of organizations (public or private, bureaucracies or start-ups) increasingly trace work or activity of their members, under the guise of diverse objectives; either supporting productivity, security, or resilience (Meijer et al., 2021; Flyverbom & Murray, 2018). Data can be collected both manually or automatically through the usage of the different devices and IT systems that equip the work or the activity. These traces (of the use of an application, of the access and edition of data) are often used to inform metrics, or to produce analytics (like activity dashboards), that are increasingly sophisticated, and therefore also support more and more granular ways of monitoring, evaluating, and improving business processes, as well as assessing employees’ productivity. 

If the digitization of work has expanded the possibility to collect traces of activities, AI techniques now extend the potential for analyzing this large amount of data. However, it remains difficult to make sense of the data that is collected and analyzed by this AI. As (Koesten et al. 2021) say: “while sensemaking of textual information has been well-explored, there is a relative gap in research that aims to understand the strategies involved in making sense of data". Indeed, human work is needed to tune algorithms, and to be able to integrate AI into real-world systems (Fiebrink & Gillies, 2018), which finally ends up increasing the cognitive overload of the workers. 

What is often highlighted is the harm that these techniques to collect and analyze data at work can cause to the workers. For instance, Levy (2022) has explored how technology (sensors, cameras, GPS systems, and on-board computers) is increasingly used to monitor truck drivers in the United States. She shows how the various surveillance technologies that are used to monitor and measure drivers’ performance further reduce their autonomy and increase the risk of penalties for minor errors. 

However, these issues could be addressed in another way; in a context where organizations will increasingly use AI, one can indeed discuss how and under which conditions the collection and the analysis of data (traces of the activity of workers) could be rather used to improve the quality of life at work (QLW), in particular by reducing their information, cognitive, and communication overload (Cicourel, 2004; Mark, 2003; Wilson, 2001). 

Actually, in a context where the development of AI increases the processing capacities of this data tenfold, it is urgent to consider uses that are not only related to control and a logic of increased performance, but also daily uses that make data meaningful and interpretable by reducing uncertainty, equivocality and supporting organizing processes (Weick, 1995). We still know very little about how users interpret usage data in real work settings; what concerns, or hopes, and forms of trust their place with usage data entail, and  how these are used to support daily practices at work. We need to examine usage data in mundane everyday working to understand how people experience working with usage data, and how in proceeding through their daily activities they take advantage of data to support collective processes (Pink et al., 2017). This involves specific methods to understand how people work on and with usage data (e.g. Kristiansen et al 2018). It is indeed important to look at how AI reconfigures work practices by producing analytics, not only looking at the technology’s potential capacity, but also on the labor of integration that humans must accomplish to correct errors or to allow a better integration of the technologies in their workplace practices (Mateescu & Elish, 2019). Employees must indeed interact, collaborate with, and integrate data and their analysis generated by AI systems into their work activities (Faraj et al., 2018; Jarrahi, 2018). In other words, to successfully integrate AI into the organization (not harming the workers), we need to consider not only its technical aspects, but also the human (“human infrastructure”, Mateescu & Elish, 2019) and social aspects ("social interoperability”, Grosjean, 2019 and “data valences”, Fiore-Gartland & Neff, 2015). It is time to explore in more detail the synergies, the forms of collaboration that can take place between human workers and AI in the workplace (Seeber et al., 2020; Flygge et al 2021, Saxena et al 2021)), and then to incorporate this knowledge into the design of socio-technical systems that support the visualization and the analysis of data collected at work (Makatius et al., 2020; Bader & Kaiser, 2019), and therefore help collectively making decisions on how to evolve for a better QLW (Paschkewitz & Patt, 2020). 

The question is then also to discuss how workers can negotiate the collection and the analysis of data, and how they can use this collected and analyzed data to reflect on their activity, both at the individual and the collective levels. These reflections could lead to collectively defining norms for QLW. We can envision that there is a heterogeneity of the employees' representations of the practices they consider problematic regarding QLW. In this context, how could communication conventions be developed within an organization and how could this collective elaboration be supported? This last question raises the issue of participatory designing systems for collecting, analyzing, and reflecting upon data at work, which is related to the conditions of appropriation of AI-based technologies: Does the possibility of "seeing in action" and revising the collection and the analysis could contribute to the transparency and the appropriation of these technological opportunities? 

References

• Bader, V., & Kaiser, S. (2019). Algorithmic decision-making? The user interface and its role for human involvement in decisions supported by artificial intelligence. Organization, 26(5), 655–672. https://doi.org/10.1177/1350508419855714

• Cicourel, A. V. (2004), Cognitive Overload and Communication in Two Healthcare Settings, Communication and Medicine, 1, 35-44.

• Faraj, S., Pachidi, S., & Sayegh, K. (2018). Working and organizing in the age of the learning algorithm. Information and Organization, 28 (1), 62–70.

• Fiebrink, R., Gillies, M. (2018). Introduction to the special Issue on Human-centered Machine Learning, ACM Trans, Interactive Intelligent Systems, 8(2), Article 7 (July 2018). 

• Fiore-Gartland, B., & Neff, G. (2015). Communication, mediation, and the expectations of data: Data valences across health and wellness communities. International Journal of Communication, 9, p. 19.

• Flyverbom, M., & Murray, J. (2018). Datastructuring—Organizing and curating digital traces into action. Big Data & Society, 5(2), 2053951718799114.

• Flügge, A. A,, Hildebrandt, T., & Møller, N. H. (2021). Street-level algorithms and AI in bureaucratic decision-making: A caseworker perspective. Proceedings of the ACM on Human-Computer Interaction, 5(CSCW1), 1-23.

• Grosjean, S. (2019). L’interopérabilité sociale de l’IA en santé : Un enjeu pour le design d’algorithmes situés dans des pratiques. Revue Française des Sciences de l’Information et de la Communication (RFSIC), Numéro 17. URL : http://journals.openedition.org/rfsic/6138 

• Jarrahi, M. H. (2018). Artificial intelligence and the future of work: Human-AI symbiosis in organizational decision-making. Business Horizons, 61(4), 577–586. 

• Koesten, L., Gregory, K., Groth, P., & Simperl, E. (2021). Talking datasets–understanding data sensemaking behaviours. International journal of human-computer studies, 146, 102562.

• Kristiansen, K. H., Valeur-Meller, M. A., Dombrowski, L., & Holten Moller, N. L. (2018, April). Accountability in the blue-collar data-driven workplace. In Proceedings of the 2018 CHI conference on human factors in computing systems (pp. 1-12).

• Levy, K. (2022). Data driven: truckers, technology, and the new workplace surveillance: Princeton University Press.

• Mark, G. (2023). Attention Span. A Groundbreaking Way to Restore Balance, Happiness and Productivity: Harper Collins Publishers.

• Mateescu, A., & Elish, M.C. (2019). AI in Context: The Labor of Integrating New Technologies (New York: Data & Society Research Institute, 2019), https://datasociety.net/wp-content/uploads/2019/01/DataandSociety_AIinContext.pdf.

• Makarius, E. E., Mukherjee, D., Fox, J. D., & Fox, A. K. (2020). Rising with the machines: A sociotechnical framework for bringing artificial intelligence into organization. Journal of Business Research, 120, 262-273.

• Meijer, A., Lorenz, L., & Wessels, M. (2021). Algorithmization of bureaucratic organizations: Using a practice lens to study how context shapes predictive policing systems. Public Administration Review, 81(5), 837–846.

• Paschkewitz, J., & Patt, D. (2020). Can AI Make Your Job More Interesting? Issues in Science and Technology, 37(1), 74–78.

• Pink S, Sumartojo S, Lupton D, et al. (2017) Mundane data:The routines, contingencies and accomplishments of digitalliving. Big Data and Society 4(1). Available at: http://journals.sagepub.com/doi/abs/10.1177/2053951717700924

• Saxena, D., Badillo-Urquiola, K., Wisniewski, P. J., & Guha, S. (2021). A framework of high-stakes algorithmic decision-making for the public sector developed through a case study of child-welfare. Proceedings of the ACM on Human-Computer Interaction, 5(CSCW2), 1-41. DOI: https://doi.org/10.1145/3476089 

• Seeber, I., Bittner, E., Briggs, R. O., de Vreede, T., De Vreede, G. J., Elkins, A., ... Schwabe, G. (2020). Machines as teammates: A research agenda on AI in team collaboration. Information & Management, 57(2), Article 103174. 

• Weick, K.E. (1995). Sensemaking in Organisations. Thousand Oaks: Sage, 3th edition.

• Wilson, T. D. (2001), Information Overload: Implications for Healthcare Services, Health Informatics Journal, 7, 112-117.