Introduction

A key insight from Grunwald’s critique of traditional, consequentialist technology assessment (TA), is that faced with the openness of the future regarding new and emerging sciences and technologies (NEST) existing merely as visions, expectations or imaginaries, TA must avoid “‘closing down’ the future space of options without sufficient evidence and thereby becoming blind to alternatives or spaces for shaping future developments in other directions” (Grunwald 2020, p. 98). The societal meaning attached to a NEST should be TA’s main object and he thus proposes a hermeneutic extension to TA to analyze contested societal meanings and their debates (Grunwald 2020). But how can alternative societal meanings emerge? Widening participation for shaping future NEST pathways faces the issue that anticipation operates within the sociotechnical context from which it emerges and thus structurally privileges certain imagined futures over others. For anticipation to help democratize future-making, its interventive processes must try identifying and disrupting its enframing context (Urueña 2022).

The projected technology in this use case falls into the NEST for health category. The project conducted exploratory research for not-yet-existing digital psychiatry technology, to which the scientific team attached a societal meaning driving the research’s direction, but which potential deployment and application in a clinical setting remained abstract. Following Grunwald and Urueña, the social scientist advising the project on social and ethical questions (Aicardi) aimed to foster alternative sociotechnical visions for it through the collaboration with a speculative fiction author and practitioner of applied science fiction (Oram). Applied science fiction (SF) is “science fiction that is trying to do something, to not only glimpse but also shape the future” (Walton and Levontin 2023, p. 3). It is emerging as a loosely defined but growing multidisciplinary field that uses SF to think about, build capacity for, and try to weigh collectively on pathways to the future – notably invested by the Applied African Speculative Fiction project and ASU Center for Science & the Imagination’s Applied Sci-Fi Project. This research article analyzes and reflects on the use of applied SF for anticipating sociotechnical challenges, and associated ethical concerns, of future digital health technologies. It details a specific use case of applied SF, but builds on the knowledge and practice developed by Aicardi and Oram across many collaborative projects since 2016.

The use case follows a 3-year project (2021–2024) piloting the development of a technology automating the analysis of ‘expressed emotion’ in parental speech using the Five Minute Speech Sample (FMSS) method, with the goal of improving the prediction of youth mental health problems – outlined in Mirheidari et al. (2024). The FMSS method, originating in 1970s’ adult psychiatry, has been used in developmental research with adults and young children since the 1990s to capture the emotional context of the parent-child relationship and predict their potential future youth mental health issues. An interviewer asks a parent to talk about their child for five minutes, then a specially trained human rater codes the speech recording for expressed emotion, which is the type and intensity of emotions a parent expresses about their child (Sher-Censor 2015). However, the resource intensity of coding has prevented translating it for routine assessment in clinics. The project aimed to evaluate if a tool driven by artificial intelligence (AI) could be built, that could transcribe and rate the parents’ speech samples at least as well as humans and in near-real time, to help clinicians identify families who may benefit from support to prevent children developing mental health problems.

The project aimed to evaluate if an AI-driven tool could be built, that could transcribe and rate the parents’ speech samples at least as well as humans and in near-real time.

The multidisciplinary scientific team, with research and clinical backgrounds in child and adolescent psychiatry, health informatics, social psychology (Fisher), AI, speech processing, linguistics (Firth), counted eight members overall. In addition, Aicardi was a co-investigator in the two successive grants funding the project, overseeing stakeholder workshops, the creative writing collaboration, and advising on ethical and social issues; and Oram was a freelance contractor in both rounds of funding. His role was to create two bespoke SF short stories, eventually published as A Mother’s Nightmare and Standard Deviations (Firth et al. 2025, supporting information; Oram 2023), about possible near-futures where the projected technology is embedded into society. The stories’ creation required a first round of activities with the scientific team and with various stakeholders as Patient and Public Involvement and Engagement (PPIE). The stories were then used as springboards for discussion in further PPIE activities, with findings fed back to the researchers, to try and be attentive to diverse perspectives for better anticipating the sociotechnical and ethical concerns of implementing the proposed technology in the real world. The paper details the methodological process used by Aicardi and Oram in the applied SF approach designed for the project, first, to create the stories and second, to organize their subsequent afterlife. Finally, we discuss key findings and insights from this 3‑year experience. Links to events and organizations mentioned in the paper, and more, are listed in Oram (2026).

Applied science fiction development phase

The practice of applied SF developed by Oram and Aicardi covers the two distinct lives of a story – development and (post-)release. In both development and reception, applied SF inserts external expertise broadly construed. The development phase starts by gathering as many diverse perspectives as possible (given constraints of time and resources) about the object of the applied SF exercise. In the present case, the development phase of Oram’s two short stories involved consultation with the project’s scientific team, along with external clinicians and lived-experienced experts (young people with mental health issues), who contributed diverse perspectives to the writing process.

A first step was understanding not just the basics of the projected technology but more importantly, which specific societal meaning the scientific team (a multidisciplinary group of clinical and research scientists and technologists) collectively attached to it. To this end, Aicardi and Oram jointly conducted six semi-structured interviews with its members, who provided different disciplinary angles. The interviews were deliberately hybrid. A first part covered typical social scientific questions about the interviewee’s research background, role in the project, expectations for success, potential issues and barriers identified. Then, Oram presented a set of possible characters for the stories: a professional (e.g. an educational psychologist); a grand-father (bringing in a male perspective otherwise missing, as FMSS recordings were only of mothers), his daughter (coping as a young mother with the outcome of her assessment as a child), and his grandchild; a possible employer of a person assessed as a child (gaining access to the data); another institutional figure (e.g. police, judge). The aim was to see how the interviewees responded to them, how they imagined these characters may be affected by the projected technology if it was successfully deployed according to their definition of success, and whether they had suggestions for other characters and roles. In this manner, the interviews collected data that could help surface the collective societal meaning assigned by the scientific team to the project while also gathering their individual responses when asked to consider the perspective of fictitious characters exemplifying certain social roles. The collective societal meaning that emerged was, broadly, one of child psychiatry in crisis, short on time, resources and personnel, where triage is often decided on subjective grounds due to the lack of agreed diagnostics standards as much as the clinicians’ personal biases, and where the projected technology may not be ideal but would be welcome if it helped making things even only slightly better. It cannot be pinpointed when this specific sociotechnical vision became clear to Oram, and how and when it influenced his writing, but retrospectively, there are traces of it in both stories.

Three creative workshops followed these interviews: two with clinicians in child and adolescent mental health; one with young people keen to be involved in mental health research from the McPin Foundation Young People’s Network. Oram was then in the process of developing the short stories, at the stage of framing the two main ideas he had settled on: one centered on a clinician’s practice, one on a big family gathering. The clinicians’ workshops were geared towards generating ideas for the clinician-centered story and consisted of mixed activities: starting from selected newspaper articles, thinking of the use of AI for mental health diagnosis/treatment and for predicting if a child might develop future mental health issues; then, more creatively, trying to imagine what would be different (for good or bad) ten to twenty years into the future, grounding their imagination in a day from their working life; finally, after introducing the FMSS automation project, returning to the previous futuring exercise to think about how this technology may be integrated in the future they had in mind, how it could be used in practice, who could be the people most affected by it and how. The workshop with the McPin group was a creative writing activity, using the story set-up prompt of a big family gathering for a birthday ten to twenty years in the future, where a younger family member applies for a job in the police force and is asked if an automated system has ever diagnosed them as likely to develop mental health issues. The participants went through different steps in the creative process and shared what they produced along the way – where their story took place, who were the characters and what happened to them, what was the mood they were trying to set, and what key message they would want to carry across.

Rip and Kulve (2008) argue that constructive TA scenarios can serve as ‘useful fictions’ for studying the implications of promised but not-yet-existing NEST. But Oram’s development of applied SF short stories is a fundamentally different process from that of developing sociotechnical scenarios (Hofman et al. 2004). First, contrary to most qualitative social scientific approaches, the data collection was designed to widen the perspectives elicited by deliberately appealing to creative imagining, be it in the interviews through the possible story characters or in the creative workshops. Then, Oram’s way to process the data collected was, again, not to analyze it according to proven qualitative methods that would return trends and drivers for building scenarios in a structured social scientific way. Instead, he has found that it is snippets here and there, in the plethora of material collected, that create the hooks on which his stories rest. In A Mother’s Nightmare, for instance, the ambivalent relationship of the clinician to the AI-powered assessment tool grew on an aggregate of multiple snippets from the scientists’ interviews and the clinicians’ creative workshops. In Standard Deviations, a snippet from a scientist’s interview provided the feelings of strangeness and annoyance that a young adult may experience at finding their parents had given speech samples in their childhood for an automated assessment of their future mental health; and the creative workshop with the McPin group produced the counterintuitive idea that in a favorable policy context, a predictive mental health diagnostic could discriminate positively against usually disenfranchised groups, by making them more likely to get a job if they have a positive diagnostic when coming from certain minority groups.

Thus, it is not any particular group of people providing the snippets, it is rather that the richer the material collected, the more of these crop up and the more diverse they are, reflecting different and at times clashing views that get woven together through the storyteller’s art into complex narratives that hover between dystopic and utopic. In other projects, Aicardi and Oram have enrolled more speculative fiction authors, who would each digest the data provided in their own way, thus multiplying the variety of alternative visions coming through the stories produced.

A key aspect of the development phase is that it respects the slow time of the creative process. Here, the data gathering and the stories’ creation spread over nearly a year. In addition, the stories followed a few simple rules born from Aicardi and Oram’s previous experience: They were no longer than ten minutes, to be short enough to be read in public; the scientific team were invited to check their scientific plausibility, but Oram retained full editorial authority; finally, the stories were not overly dystopic or utopic, for two main reasons: First, the stories should not be so off putting that the scientists involved would refuse to share a stage with the author at a public reading; second, grey areas are more interestingly complex to explore from Oram’s literary standpoint as much as Aicardi’s social scientific one.

Applied science fiction post-release phase

Applied SF stories’ second life starts with their release into the world, as for any stories: when they are read, passed on, interpreted, meaning different things to different people. But in applied SF, this second life is harnessed so to speak, through carefully curated activities that can draw on multiple forms of expertise. In this case, there were two main types of activities: Open public events, where Oram read the stories to kickstart discussions with the audience; and PPIE workshops with specific stakeholders’ groups.

A first public event was organized by the project in June 2022 in London, with a panel of Oram and members of the scientific team representing different clinical/research fields. The much-reduced number of participants (about 15), due to a train strike, proved beneficial in the scientists’ view as it fostered an informal, interactive dynamic between panel and public compared to larger public engagement events. The highly engaged audience pursued follow-up questions, making for deeper interactions. The quality of informal but deep engagement also compared positively to the formality of institutionalized patient advisory boards. Oram took part on his own in many more public events at which he read one or both stories, collecting more pertinent insights that he fed back to the scientists and to inform future applied SF projects. Notably, using the stories in different contexts and locations (e.g. various SF conventions, an online Digital Futures conference organized by the Goethe Institute Indonesia, AI-ethics workshops with students and teachers at the University of Zurich) highlighted cultural differences regarding the role of mothers: While certain groups understood the mother’s stance in A Mother’s Nightmare, others judged her negatively, arguing that a (good) mother would never act in this way.

Besides, we ran four PPIE workshops in 2023–2024: with Lambeth Young Carers, with a group of young people with experience of mental health problems from the NIHR Maudsley BRC’s Adolescent Mental Health Advisory Group, and two with parents (some autistic, some not) of autistic children from Autistica. The workshops were designed as collective readings of the two stories, circulated in advance if requested by the participants, to kick off discussions.

Previous PPIE workshops during the stories’ development phase involved traditional stakeholders – clinicians and an established young people’s group. Aicardi and Oram’s applied SF practice has repeatedly shown that drawing in more marginalized or vulnerable groups into futuring exercises, at this first stage, is extremely difficult, for various reasons falling outside the scope of this article – an issue they are still researching, empirically and theoretically. However, the present project showed that with applied SF, the post-release phase is an opportunity for expanding PPIE activities to social groups otherwise overlooked or difficult to connect with. Autistica, UK’s leading autism research and campaigning charity, may be a traditional PPIE actor but they represent a historically under-heard group. The project did not have an explicit focus on autism, but Autistica agreed to our collective reading workshop proposal and were highly engaged participants. As for the Carers’ Hub Lambeth, it is not a patient support or advisory group. The Young Carers group’s remit is to support and organize activities for under-21 youths in the Borough of Lambeth who help care for a family member or friend with an illness or condition. Yet the young carers’ experience of caring for those meant to be their carers provided a unique perspective on the use of expressed emotion in parental speech to predict youths’ future mental health. We thus organized the workshop as a reading group activity offer for the half-term school break.

In this post-release round of PPIE workshops, Oram’s stories proved a useful medium to engage with different social groups in a variety of settings across the UK and abroad. They elicited narratives from utterly different contexts than the ones imagined originally. They were especially valuable to enter into conversation with more vulnerable individuals. The stories’ characters lived through situations that participants related to, making the projected technology far more concrete, and evoked strong feelings, positive and negative. We found that, by talking about and through the characters and situations encountered in the stories, some participants seemed less reluctant to relate their own experience, while others could put it at a distance and talk about seemingly close-to-home topics while protecting themselves. For instance, a participant suggested that what a parent says about their child may not be trustworthy because how a child behaves at home can be the opposite of at school, e.g. silent at home but loud and disruptive at school.

Discussion

The outcomes of the project’s PPIE activities are synthesized and discussed in Firth et al. (2025), we thus focus on the outcomes that the applied SF process had for the scientific team. First, they were pleasantly surprised at how well it worked. Second, they were impressed by the richness, quality and nuances of the points and questions raised at the June 2022 public engagement event and at PPIE workshops. The positive conclusion was that “[t]he nuance and breadth of participants’ insights generated by the stories, including on technical aspects of the tool, indicates that factual or technical descriptions of a technology are not the only types of information that breed understanding” (Firth et al. 2025). Finally, the richness and pertinence of the insights from the PPIE activities fostered deeply reflexive discussions between the scientists. For instance, it strengthened the view that the development of their projected AI tool should not be rushed, and that it should stay within the National Health Service and not involve private interests from the tech industry. Another unexpected discussion was about whether mental health outcomes should be predicted at all, AI or not. This questioned the assumption, widely held in clinical research, that prediction is the best form of prevention.

Bringing together the PPIE and scientific team’s outcomes, a key finding, we argue, concerns futures literacy in the context of TA and anticipatory governance (Miller 2018). For Miller (2018, pp. 1–13), futures literacy is not about accuracy of prediction, perfect preparation and planning. It is rather the capacity “to reveal the anticipatory assumptions that determine why and how futures are imagined” (p. 5) and to deliberately, reflexively, ‘use the future’ – different kinds of futures – in various ways, to become active ‘designers of meaning.’ Unfortunately, central questions (What is the future? What kind of future to anticipate? Which methods to think about different kinds of future?) are seldom posed, which may be due in part to futures illiteracy (Mangnus et al. 2021; Miller 2018, pp. 1–13), because “the ability to imagine and strive toward a life or society different from one’s present situation, is typically an implicit phenomenon” (Mangnus et al. 2021).

Böhnert (2025) suggests that “futures are materially-discursively enacted in social practices, giving rise to a multiplicity of temporalities and ultimately constituting futures as cultural facts. […F]utures do not exist as discrete endpoints but emerge relationally” (p. 19); yet “despite the entanglement of present and future, dominant frameworks often assume a linear temporality” (p. 16). He analyzes the challenges this poses for anticipatory governance, with a focus on long-term governance, and makes recommendations to address these (Böhnert 2025). He could very well have formulated his diagnostic about dominant frameworks in terms of futures illiteracy. Considerations of futures literacy, we argue, should underpin all efforts of participatory anticipation in NEST research, for the different stakeholders’ groups and members of the public as well as for the scientists and technologists involved. It is especially important if we embrace hermeneutic, or constructive (Douma et al. 2007; Rip and Kulve 2008), TA approaches.

Applied science fiction can help develop futures literacy capacities for all involved.

Applied SF, through the stories’ creation process and their subsequent uses, can help develop such futures literacy capacities for all involved. In our case, it equipped different groups of stakeholders and audiences (with some that find it difficult to get involved), and the scientific team, with tools to identify and think about different kinds of sociotechnical futures. For the scientific team especially, stepping outside the enframing sociotechnical future they had imagined for the projected technology and reflexively questioning taken-for-granted assumptions exemplified (steps in) the journey to futures literacy. Moreover, following Mangnus et al. (2021) in their analysis of different categories of futures approaches, they developed futures literacy in two distinct approaches: experimental (aimed at the collective co-creation of new futures) and critical (similarly constructivist but “scrutinizing imaginaries in order to shed light on their political implications”). These two kinds of futures approaches, underrepresented in anticipatory practices (Mangnus et al. 2021), are highly relevant to hermeneutic TA.

Conclusion

Through a bespoke applied SF process, we have defined, analyzed, and reflected on the use of applied SF for anticipating the sociotechnical and ethical challenges of future digital mental health technologies. Key findings concern the impact on the scientific team, who became aware and attentive to alternative societal meanings of their projected technology, and questioned taken-for-granted assumptions. Developing future-oriented critical reflexivity is part of futures literacy, an essential but often overlooked underpinning of anticipatory practices in TA for NEST.

This applied SF project, the longest ran to date by Aicardi and Oram, brought positive elements, theoretical and practical, to their ongoing experimentation with the use of applied SF in participatory sociotechnical foresight for NEST and opened up directions for future investigations. It was their first foray in using applied SF in health research with a formal PPIE component. The outcomes of the PPIE activities, and their reception by the project’s scientific team, show that such an applied SF approach may help address some shortcomings and main critiques levelled at PPIE in its current forms (Mayeur et al. 2025; Montenegro and Green 2025). Another promising area for further investigation relates to interdisciplinary studies. The findings indicated that applied SF may help multidisciplinary teams to create joint understandings of their exploratory research and to develop their capacity for communicating across disciplinary divides.