There's No App for Healthcare Staff Shortage

Abstract

Governments worldwide are facing significant healthcare staff shortages, driven in part by aging populations with greater health and long-term care needs. Digital technologies, including robotics and artificial intelligence (AI) supported systems, are frequently presented as the primary, or even sole, sollution to this challenge. However, such technological solutionism is misguided. Technologies do not simply substitute for missing (or migrating) workers as labour. Rather, technologies enter into existing care relations and reshape care practices from within. As a result, they do not reduce workload so much as transform the tasks and routines that constitute healthcare work. Beyond the general value of ethnographic method, anthropologists are especially well positioned to illuminate what typically remains invisible in studies of technologies within healthcare organization.

On a Thursday afternoon, I sat in a mid-sized, windowless conference room, half-full and half-empty – a fitting image for a conversation about technology that rarely admits a middle ground. The conference brought together participants from Dutch academic institutions to discuss public values for (not in) a digital society. This particular panel was on health and wellbeing, and was concerned with how to overcome challenges in healthcare through digital and increasingly artificial intelligence (AI) supported solutions. A team of researchers presented their digital solution: an application designed to support mental health provision and help fill the gap left by insufficient staff. Years in the making, the application still struggled with a number of unresolved issues. It was not yet fit for implementation. The bottom line, as the presenters put it, was more funding.

The presentation was followed a discussion facilitated by another digital solution, an online questionnaire, accessible via QR code, that collected audience responses in real time. As audience members struggled with their wi-fi connection, capturing the code, and finding their way through the online questionnaire, time was ticking. But once our responses started appearing on the screen before us, the word I had entered stood out enough for the panel chair to pick it out: Why? Why was entering a response through a QR code preferable to simply asking the audience to speak? What were the underlying reasons for producing these applications and digital solutions that were expensive and took years of development for little demonstrated return on investment? To me, why? was not a rhetorical question: it was one that ought to precede any investment.

The answer was grounded in the story of potentiality. The presenter responded, almost with anger, that technological innovation was simply essential: there weren't enough people, she insisted, to attend to the demand for mental health support. The app might not be there yet, but it is vital and inevitable.

The answer was telling because of what it revealed about the terms of the debate. In her response, the presenter treated the staffing shortage as a fixed condition and technology as the only available variable, as if the choice were innovation or nothing at all. This is the logic of technological solutionism: that a persistent, structural problem (too few healthcare workers, too much demand) can be resolved by building a better application, rather than by asking why the problem exists, and whether an app or another kind of technological or non-technological solution is the right kind of answer to it. 

Robotic Replacements

Over years of attending academic and industry events on digital technologies in healthcare, now a fixture of the Dutch conference calendar, I have repeatedly heard the same narrative: the need for healthcare is rising, partly due to aging populations with greater health and long-term care needs, but the number of healthcare staff is insufficient. We live in times of great technological advancements, so the key to this "care crisis" must be technological innovation.

This narrative is not unique to the Netherlands. In 2022, the World Health Organization (WHO) published a media release warning that health and care workforce gaps across the European and Central Asian Region had become a "ticking timebomb," with countries facing a serious and worsening shortage of doctors, nurses, and other care workers. The release noted that in some countries, as much as 40% of doctors were nearing retirement age. Additionally, in the midst of the Covid pandemic, WHO had received reports of as many as nine out of ten nurses declaring their intent to quit their jobs. The piece ended with ten proposed actions to strengthen the health and care workforce. But the solution most consistently on offer is strikingly singular: technological innovation and the expansion of digital technologies, such as robots.

Europe is not the first to walk this path. One of the most rapidly aging countries in the world, Japan, has strategically pursued a policy that privileges automation over international labor migration. In her ethnography of humanoid robots in Japan, Jennifer Robertson (2018, 123) observes that whereas European and North American imaginaries tend to cast robots as a threat – even cute care robots running amok – robot caregivers in Japan are often framed as preferable to foreign nurses, valued for their capacity to cultivate what Robertson calls "ontological security" (2018, 144). The robot caregiver, in other words, is never only a technological solution to labor shortage, but also a way of deciding who should be permitted to do that labor and on what terms.

When robots and other AI-powered technologies are positioned as a fix for healthcare labour crisis, it matters who they are imagined to replace. These are commonly low- and middle-skilled carers – disproportionately women and often migrants from formerly colonized countries (Wright 2023). The premise underlying most digital health technologies is a division of labor: some care tasks cannot be substituted by technology, but those by low- and middle-skilled carers can. Should this vision materialize, those workers are likely to see their labor most devalued (Breuer and Müller 2024, 959). The technoliberal project of building robots to replace human workers is therefore far from a neutral technical undertaking. It is embedded within, and simultaneously obscures, the uneven gendered and racial structure of care labor itself.

The idea that technology, including robotics, could, and even should, substitute migrant labourers is not confined to the healthcare sector. At a conference on technological refusal, organized at the University of Maastricht in February 2026, Koen Beumer and Lilya Khachatryan described a strikingly similar logic at play in the field of agricultural robotics, whereby automation was imagined not simply as a response to labour shortages but to reduce reliance on migrant workers specifically. The care sector is thus not an isolated case, but one instance of a broader pattern in which labour traditionally performed by migrants is the labour targeted for replacement by machines.

Hacking productivity

Investment in care robotics has been substantial since the mid-2000s: Japan devoted at least €275 (about $315) million and the European Union (EU) at least €235 (about $269) million in combined public and private funding (Wright 2023, 6-7). Nevertheless, care robots have not become commonly used anywhere, not even in Japan. Instead, many of these robots, including the humanoid robot Pepper, marketed for companionship in healthcare and elder care facilities, have since been discontinued.

A major reason these technologies failed to find a sustainable market is that although designed to watch over care home residents, the devices "needed to be watched over by care workers" (Wright 2023, 127). I encountered a similar dynamic in my own fieldwork on animal-shaped robots in Dutch elder care institutions. The care workers I encountered immediately linked the introduction of technologies in elder care as an attempt to counter labour shortages. They also immediately laughed the idea off, confident that no robot cat would ever replace a missing pair of human hands. Whatever their form, robots are clearly not going to replace people in health and long-term care any time soon. They do not function effectively in daily care practices and require significant effort to keep them afloat, adding to the workload of already strained staff (Chevallier 2022).

If robots have failed to replace people as laborers in (health)care, a related but distinct promise has taken their place: technologies not for substitution, but augmentation. Chatbots, ambient scribes and other digital tools are being designed to increase efficiency on the healthcare floor. The focus is not on direct care as much as on supportive tasks, such as sample transportation within hospitals and administration. Why spend precious time taking notes and writing reports, if an AI system can listen in and transcribe the doctor-patient conversation in real time? Even privacy concerns around such practices have been tackled, for example, by ensuring that the data are stored locally rather than in a cloud, and for a limited time only.

Recently, I participated in a presentation of a chatbot, designed for an intake with patients, which is expected to save a couple of hours of conversation. At the end of such an intake, the patient would meet a human again, so ensuring the reliability of data collection and safety for the patient when deciding on the next step of their personalized care.

The underlying problem with such attempts to mobilize technological innovation to improve efficiency in healthcare is that they operate on the logic of the market, while care follows a fundamentally different logic. As Annemarie Mol (2008) argues, health and long-term care are not about delivering products, but about sustaining processes – the process of becoming healthy again, or of living a good life with a chronic illness. When ministries of health, in partnership with technological industries, respond to healthcare staff shortages and rising expenditure by turning to digital technologies, they attempt to translate the logic of care into the logic of the market, as something that can be streamlined, counted and made more efficient and productive.

These two logics do not translate cleanly, and the equation "people + technologies = higher productivity in healthcare" does not hold. Technologies are not 'just tools,' but are fundamentally relational. They become (dys)functional, or function differently than originally intended, in relation to people and other objects, institutions, ideologies, words, and so on. Ethnographers have long shown that inserting technologies in (health)care practices does not reduce the workload – and therefore the need for healthcare staff – but rather, changes the daily routines of care work. This is the case for robots (Wright 2023) as well as for digital systems such as telemonitoring (Pols 2012). Indeed, any technology, from the simplest singing digital ball recently presented to me as a potential rehabilitation device, to everyday, easily available and accessible devices such as smartphones, requires work for their integration and mainentance if they are to support health and care (Ahlin 2023).

Attending to the Unattended

This is not to argue that digital technologies are impossible to integrate into healthcare, or that they cannot improve it. If technological solutionism is not the answer to healthcare's crises, what would a genuine alternative look like and how could anthropology help?

Any technological investment and innovation must first answer a simple question: why? Why, for example, is a €6000 digital ball more efficient for rehabilitation therapy than €1 soft ball? As Tamar Sharon has recently noted at several academic events, it is high time for serious financial reckoning with technological investments in healthcare. How much has been spent so far, and what, concretely, has it yielded? Without honest answers to this question, it is difficult to justify future investments.

Asking 'why' is not only a financial question, but a conceptual one. It requires that we keep the repertoire of possible solutions as wide as possible and resist the assumption that newer or more complex technology is always better. Launched in 2004, the seal-shaped social robot Paro has deliberately not been updated with the latest AI technologies, and yet, unlike some other social robots, it remains on the market (Shibata 2023). Even simpler and considerably cheaper robotic cats and dogs are still used in Dutch care institutions where they were first introduced almost a decade ago. They do not replace care workers, but care staff do acknowledge their positive impact on specific residents when used thoughtfully and at the right moment. Sometimes, the least technologically sophisticated robot is the one that works best in practice.

A wider repertoire of solutions includes the option of setting technology aside. Non-technological solutions – such as aligning healthcare education and professional development with actual needs, as WHO (2022) proposed – deserve equal attention, even though they rarely generate the same excitement as a new device or app. This asymmetry played out at a seminar talk on a healthcare chatbot where a speaker displayed a slide stating that AI was "the only solution" to the healthcare crisis. When I challenged that claim, he backtracked: that framing, he said, had been a mistake; AI was not the only possibility after all. Yet in the same breath, he added confidently that national healthcare budgets were not about to increase any time soon. His response is telling. It reveals what governments believe about healthcare needs, what they are willing to fund, and what that funding pattern says about their underlying values.

This asymmetry, between what is funded and what is needed, requires a particular kind of attention. This is where anthropologists have a crucial role to play in technological innovation for healthcare. Other disciplines have laid claim to ethnography, or at least to observation, as a method for participatory design. But anthropologists bring more than that: our training and practice keeps us open to serendipity and attentiveness to what routinely goes unnoticed. We notice, for instance, that human-machine interaction is neither purely dialectical nor simply bi-directional, but collective. For a technology to be adopted successfully, it is not enough to understand the relationship between a machine and its user; "arbitrary care collectives" which also include peer-patients may prove just as decisive (Ahlin And Mann 2025).

Beyond the broader social, historical, political and economic forces anthropologists are trained to trace, we also attend to what happens in practices, spaces and silences where seemingly nothing at all is happening. The time spent talking with a patient is not only to extract information about their care preferences, but to build trust. The time a nurse spends carrying tissue samples between hospital departments may not be time lost, but a moment of mental and emotional recalibration. The time a doctor spends writing notes by hand may be a way of actively processing information rather than merely recording it.

If healthcare is to genuinely benefit from technologies, anthropologists are needed to attend to what most often goes unattended. Perhaps that may help to slow the pendulum swining frantically from technological optimism on one side to technological pessimism on the other side, and finally find the middle ground of technological realism, where certain, perhaps even not particularly expensive technologies can indeed support people in providing better (health)care for each other.

References

Ahlin, Tanja. 2023. Calling Family: Digital Technologies and the Making of Transnational Care Collectives. New Brunswick, NJ: Rutgers University Press.

Ahlin, Tanja, and Anna Mann. 2025. “Ambiguous Animals, Ambivalent Carers and Arbitrary Care Collectives: Re-Theorizing Resistance to Social Robots in Healthcare.” Social Science & Medicine 365: 117587. https://doi.org/10.1016/j.socscimed.2024.117587.

Breuer, Svenja, and Ruth Müller. 2024. “Digitalization, AI, and Robotics for Good Care and Work? German Policy Imaginaries of Healthcare Technologies.” Science and Public Policy 51 (5): 951–962. https://doi.org/10.1093/scipol/scae036.

Chevallier, Martin. 2023. “Staging Paro: The Care of Making Robot(s) Care.” Social Studies of Science 53 (5): 635–659. https://doi.org/10.1177/03063127221126148.

Mol, Annemarie. 2008. The Logic of Care: Health and the Problem of Patient Choice (1st ed.). London: Routledge.

Pols, Jeannette. 2012. Care at a Distance. Amsterdam: Amsterdam University Press.

Robertson, Jennifer. 2018. Robo Sapiens Japanicus: Robots, Gender, Family, and the Japanese Nation. Oakland, CA: University of California Press.

Shibata, Takanori. 2023. “Paro, Interactive Therapeutic Robot That Stimulates Patients with Dementia, & Alzheimer’s.” AI for Good, July 6–7.

WHO (World Health Organization). 2022. “Ticking Timebomb: Without Immediate Action, Health and Care Workforce Gaps in the European Region Could Spell Disaster.”

Wright, James. 2023. Robots Won’t Save Japan: An Ethnography of Eldercare Automation. London: Cornell University Press.

The Version of Record of this manuscript has been published and is available in Human Organization (17 August 2026) - if citing, please use this as a reference.