Digital phenotyping – Editorial

There is an astonishing posthuman promise in digital phenotyping, as Beth Semel recently argued (Semel, 2022). The goal of digital phenotyping enthusiasts is no less than to bypass the human observer as a deeply flawed threshold of medical knowledge production. The second goal is then – ultimately – to rid the human body and mind of its frailty and to utilise technology for a ‘world without disease’ (Topol and Corr, 2019). This promissory rhetoric is not only geared towards the disruption of dated medical conventions but comes equipped with bold, revolutionary concepts. Objective knowledge, based on aggregated, automated, and sweeping data collection to deliver granular, minute, and personalised healthcare; digital phenotyping is a collection of ideas, technologies, and practices to realise a powerful and futuristic vision of a medicine far beyond human capacities. This posthuman promise might be naive and driven by an abundant positivism, but as a small movement, made up of medical researchers and digital disruptors alike, it has continuously gathered steam over the last decade. The purpose of this collection is foremost to take stock and to collect a range of critical questions for a first revision of what digital phenotyping might be and what it could potentially become. The meaning of digital phenotyping is not as well defined as the many publications in this growing body of scholarship might suggest. Some of that vagueness has been captured in the critical literature. Birk and Samuel, in their sociological analysis, have described the term recently in more general terms as an analytical concept that presumes simply that diseases and illness are by and large ‘measurable by digital devices’ (Birk and Samuel, 2020). This assumes that a person’s experience of any kind of suffering is always in one way or another expressed in the digital traces of their behaviour. The leg injury that might result in a different mobility pattern; measurable tremors in the thumb control of smartphones as a sign of Parkinson’s; sudden lack of social interaction as a sign of depression: digital phenotypes can in theory be defined for any illness and disease and captured by any of the sensors, devices, and technologies, through which humans leave digital traces. Loi, in his ethical and philosophical exploration of the digital phenotype, assumes it in more general terms to be ‘an assemblage of information in digital form, that humans produce intentionally or as a by-product of other activities, and which affects human behaviour’ (Loi, 2018). Many questions remain, not least why and how this concept seeks association with genetic terminology. What does the wholesale capturing of a human’s digital traces as phenotype imply? What does it mean to group a sheer endless range of symptoms within the paradigm of inheritable traits and how does this framing structure research on and with digital phenotypes? The phrase itself was coined by the physician Sachin Jain and colleagues at Harvard in 2015 in a letter to Nature Biotechnology. Conceptually, they conceived of digital phenotyping with reference to Richard Dawkins’ elaborations on the ‘extended phenotype’ (Jain et al., 2015; Dawkins, 1982). Not only did they see digital technologies equipped to deliver a never-before-seen mass of potentially valuable data for diagnostics and prognostics but importantly these data were produced beyond the brief and cursory encounters between patients and physicians. The full-scale exploitation of these data would enable new insight into disease expressions over a lifetime. This was not only an expansion of surveillance but would open a new paradigm of medical knowledge production: rather than just recording symptoms in a medical consultation, ‘digital phenotypes redefine disease expression in terms of the lived experience of individuals, which expands our ability to classify and understand disease’ (Jain et al., 2015). In a 2017 JAMA article, the American neuroscientist Thomas R. Insel conceptualized digital phenotyping into nothing less but a ‘New Science of Behaviour’ (Insel, 2017). Since then, the phrase has given