This article investigates the media's construction of public perceptions of future human–machine relationships related to artificial intelligence (AI) development and reflects on how such perceptions play a role in shaping strategies for the use of AI in Denmark. Through a critical discourse analysis of 253 newspaper and magazine articles published from 1956 to 2021, it shows how conflicting discursive positions are constructed, representing what I refer to as public AI imaginaries. The analysis shows that newspapers and magazines tend not to distinguish between futuristic descriptions of the human–machine relationship of AI and the human-centred principles of intelligence amplification (IA). Furthermore, it demonstrates how principles of IA are reflected in the Danish strategies for AI in practice. While the discursive ambiguity has fuelled public debate, it leaves the term AI relatively vague, thereby creating uncertainty rather than possibilities for a form of human-centered AI in empirical reality.

Recently, automated decision‐making (ADM) has been increasingly introduced in for example, the public sector potentially ensuring efficiency and more just decision‐making. The increasing use of ADM has been reflected by a growing interest by scholarly research. While initially mainly researchers within law and computer sciences engaged with ADM, there has also been a growing engagement by social science and humanities‐oriented researchers. This article traces the emergence and evolution of ADM research beyond computer sciences and engineering with a specific focus on social sciences and humanities by identifying central concerns and methods while outlining a stable baseline for future research. Based on a systematic mapping of publications, we outline the contours of ADM as an area of research engaging with an emerging empirical phenomenon. Drawing on findings from the mapping, we discuss ways ahead for ADM research as part of the subfield of digital sociology and suggest that sociological media and communication studies have a crucial role to play in developing future research avenues. Drawing on advances made in audience research, we suggest a radically contextualized and people‐centered approach to ADM. Such an approach would help to develop ADM and ground it alongside people's divergent capabilities and contextual arrangements.

Arbejde medieret gennem digitale platforme indebærer en stor mængde usynligt arbejde. Vi undersøger dette arbejde gennem et kvalitativt studie fra arbejdernes perspektiv på rengørings- og madleveringsplatforme i en overvejende dansk kontekst. Vores studie viser, at det usynlige arbejde består i følgende: Supportarbejderne, der er ansat ved platformene, understøtter arbejdet; platformsarbejderne påtager sig arbejde med de algoritmiske systemer; kunderne koordinerer med platformsarbejderne, og i tilfældet af madleveringsarbejde så koordinerer restaurantarbejderne med madbude. På denne baggrund hævder vi, modsat tidligere studier af crowdwork, at selv relativt synlige platformsarbejdere, som er involveret i rengøring og madlevering, leverer store mængder usynligt arbejde. Den såkaldte ”algoritmiske ledelse” på platformene beror altså i disse tilfælde på menneskelige skøn og vurderinger, og vi diskuterer implikationerne heraf i en bredere samfundsmæssig kontekst.