Planning cells in technology research and development – participation for a multisensory screening system in the hospital context

Abstract

Mini-public participation procedures with random selection have a long tradition in parliamentary technology assessment and acceptance research. Less documented is their potential in the area of research-based development of technical innovations. This article marks a contribution to that area and discusses the application of planning cells for a citizens’ report for the participatory development of a multisensory screening system that, in the context of a global health crisis, shall contactlessly detect characteristics of respiratory virus infections to protect critical hospital infrastructures. Randomly selected citizens and clinic personnel from the Homburg region in Saarland, Germany, deliberated on the framework conditions for the AI-based data processing and ethical aspects of the so-called VI-screen system. This evaluation demonstrates how the structural anchoring of the method in an early phase of the development process was conducive for the competent researchers to incorporate the knowledge – and discussion-based recommendations of the participants, potentially increasing the social connectivity, diffusion, and responsible application of the VI-screen system in the future. Based on a critical assessment, potential adjustments to the standard participation procedure are derived in order to increase its potential for future use in research-based technology development.

Keywords:

• mini-publics
• planning cells
• technology research and development
• hospital
• COVID-19
• participation
• ai

1. Introduction

The use of deliberative participation methods with random selection, commonly grouped under the name mini-publics (cf. Voß, Sayman, and Schritt 2021, S. 112f.), have a decades-long tradition in parliamentary technology assessment (TA) and social science acceptance research (cf. Grunwald 2022; Grunwald and Saretzki 2020). Additional, corresponding formats were being increasingly advocated and demanded not only by research sponsors, but also by researchers and the public. For example in Germany, the Federal Ministry of Education and Research (BMBF) has placed participation as a cross-cutting topic at the centre of the 2022 Science Year (BMBF 2023) and has itself invited a random selection of citizens to a Citizens’ Research Council, whose recommendations have been incorporated as the ‘social perspective in the development of the participation strategy’ (BMBF 2023b). Similarly in 2022, the leading alliance of German research organisations emphasized the importance of participation in research in a common statement as it ‘shapes the everyday lives of citizens in a variety of ways’ and ‘conversely represents a way in which social concerns can find their way into research – adding value to both science and society.’ (Allianz der Wissenschaftsorganisationen 2022). Citizens’ participation in research and development project may contribute to a diversity of perspectives; an expansion of the knowledge – and database; as well as an increased social connectivity, diffusion and application of research outcomes. It is further anticipated that citizens’ participation and relevant consideration of their perspectives may not just contribute to greater transparency and openness of research processes, but as well as to greater acceptance of science in the public sphere (cf. BMBF 2023a).

The application of mini-publics in research, however, is still a niche as overviews suggest (e.g. Sayman and Hügel 2021). Correspondingly, only a small body of academic literature exists that is documenting and discussing practical experiences. Drawing on the example of planning cells for a citizens’ report in the interdisciplinary research project ‘Integrated neurotechnological architecture for contactless screening of virus-related respiratory diseases’, short title ‘VIRUS-Screen’⁣ (hereinafter referred to as the short title ‘VI-Screen’), this article aims to contribute to the reflection of mini-public methods for the research-based development of new technology applications.

VI-Screen has been funded by the BMBF since 2021 until the end of 2023 and is carried out in cooperation with the University Hospital of Saarland (UKS), the Faculty of Medicine of the University of Saarland (UdS) and the Technical University of Berlin (TUB). The development of a monitoring system was initiated, which with increasing technological maturity should gather and evaluate various contactlessly measured data, in particular video recordings, with the help of artificial intelligence (AI) and thereby recognize characteristics of infectious respiratory diseases. This application, hereinafter referred to as the ‘VI-screen system’, may in future be able to support the health examination of patients, visitors, clinic personnel and external service providers in the hospital context. Against the background of the Covid-19 pandemic and in foresight of future health crises, this is intended to better protect critical health infrastructures.

The integration of planning cells into the research project assumed that the achievement of this goal depends not only on whether the system can actually diagnose virus-related respiratory diseases with high reliability, but also on whether such a system is accepted by those affected in its functioning as intended by the researchers. The latter cannot be guaranteed against the background of a tendency of the German population to be critical of technology in international comparison (cf. Acatech Deutsche Akademie der Technikwissenschaften 2022), as well as in light of social controversies and protests in connection with the Covid-19 pandemic regarding the proportionality of legal restrictions on democratic rights (Frinken and Landwehr 2022) and the use of surveillance technologies (Kitchin 2020).

With this, the research object may be a prime example for the development of a technologically complex and potentially controversial invention that may influence everyday lives of citizens since all persons who aim to enter a clinic would possibly be processed by such an established screening system. Participation in the research-based development phase was therefore an normative imperative (cf. Grunwald 2022) and anchored already in the application phase for the VI-Screen project. At the same time, it was assumed that the quality of the technological outcome can be increased through the use of a deliberative participation format. That is, researchers would consider the recommendations of citizens which have been developed on the basis of factually and socially sound arguments (Abels and Bora 2004, 10–11)⁣⁣ for resolving ethical and data-related issues associated with such a system.

Pointing on the funding of the project under conditions and given the research and development interests of the competent researchers involved, it should be noted that no fundamental questions were discussed in the planning cells regarding the project orientation and whether such a VI-Screen system should be developed at all (Stirling 2008). Accordingly, the mini-public participation procedure in Project VI-Screen had the following objectives: Firstly, to give citizens and hospital personnel the opportunity to gain an early insight into the research process and to contribute their perspective to the development of the technology through knowledge – and discussion-based recommendations. Secondly, to empower the researchers in the formulation of relevant research questions to be addressed by the participants in the planning cells and in the utilisation of results from the citizen’s report for research and development purposes. Based on the process experiences and results of this application example, this article should meet the need for reflection with regard to the following questions:

1. To what extent were the participation goals for both the participants and the competent researchers achieved by the planning cells and the citizens’ report in the VI-Screen research project?

2. Building on this, what adjustments to the standard participation procedure could be helpful to promote the achievement of these participation goals in comparable research-based contexts?

To answer these questions, the article first establishes a theoretical reference to the use of mini-publics in technology research. It then briefly describes the characteristics of the participation procedure planning cells for a citizen's report and introduces selected examples of this method in prior research on technology development. Drawing on the VI-Screen research project, the article then presents the practical application of the planning cells and the outcome of the citizen’s report. In the following, the research questions are discussed and recommendations for potential adaptations of the method in the particular context of research are derived.

2. Mini-publics in research

Technological progress in response to societal needs and problems describes a central paradigm of our time. However, experience shows that this approach itself creates new risks. Not least through short-termism or a disregard of forecasts, the development and deployment of new technologies produces undesirable consequences, for example for the global climate, the environment, human health and safety or societal peace. In parallel with technological change, modern societies are therefore challenged to (further) develop institutions and instruments in order to be able to identify negative potentials of new technologies at an early stage and regulate them in a sustainable manner (Renn 2008).

Accompanying research plays a central role in this respect. Its original and primary function is to advise policy makers (cf. Bechmann 1992; Joss 2002) based on mostly implicit normative concepts such as social compatibility, public welfare orientation or sustainability (Moser 2018). In the traditional practice, accompanying research was frequently limited to integrate different social perspectives and system references into the evaluation of new technological products and services. It was foremost with the establishment of participatory methods ‘which aim at the participation of laymen and/or stakeholders in processes of assessment and evaluation of technology impacts’ (Abels and Bora 2004, S. 13) that this shortcoming was addressed. Through such forms of inclusion in research processes, the responsible researchers may epistemically determine interests and values of the stakeholders involved or the parts of the population they represent in order to assess the possible path dependencies of technological research and development. From this, concrete responsibilities and calls to action are to be derived for actors responsible for the further research, design and dissemination of these technologies (ibid.).

Mini-publics in research assume thereby that citizens are willing and able to work cooperatively and on the basis of a mutual recognition of the better argument to arrive at a common set of values and a shared vision for desirable roles and utilisations of research outcomes in society. The method aims to achieve objectively and normatively correct decisions regarding the development of the research object. At the same time, the approach can foster a sense of community and responsibility among participants as they develop informed and well-considered positions based on their own experiences. Documenting the participatory process in mini-publics, for example through a citizens’ report or an advisory opinion, may enhance transparency for the broader public about the basis for the participants’ recommendations and potentially influences further developments. Additionally, participants themselves may contribute to an informed public debate about the research and its outcomes within their own communities (Einsiedel and Eastlick 2000; Smith and Wales 1999, 2000). From the perspective of participatory or deliberative theories of democracy, all these aspects – knowledge acquisition, transparency, and a strengthened public debate – can be understood as a ‘value in itself’ (Moser 2018).

In practice, however, these theoretical assumptions are challenged. That is especially because there are only a few systematic discussions on the question of how these added values can be measured and which forms of procedure and methods can actually provide such outcomes within the respective institutional research framework (Guston 1999; Mayer, de Vries, and Geurts 1995; Rowe and Frewer 2004; Zimmer 2003). With regard to the evaluations of mini-publics, a distinction should be made between an self-evaluation by the process support carrying out the participation procedure, and an independent scientific evaluation (Ayano 2021).

Yet another relevant critic in the context of participatory research through mini-publics is a potential lack of bindingness of the recommendations developed by the layperson involved by the competent researchers. To influence normative decision-making, scholars suggests a structural and instrumental integration of corresponding participation formats into the respective research process as highly relevant (Joss 1998). From the researchers’ point of view, the demand for participation has in turn been criticized to de-professionalise science and potentially curtailing the granted freedom in research (Scheuch 1986).

What is more, participation in research must not be facilitated as good democratic practice to strengthen public discourse. Rather, it can be motivated to achieve extrinsic goals. That is, for example, to increase the acceptance of the research and the exploitation potential of technical developments (cf. Thorpe and Gregory 2010). Regarding the latter, participation is not neutral to its outcome, e.g. the rejection of a technological innovation or its envisioned application scenario by involved citizens might be considered a failure by the initiators. In such settings of participation in target-oriented research and development, the exploitation of existing power differences between the facilitators, competent researcher and layperson involved must be examined, which can unfold not least in the agenda setting, in the selection of experts and participants as well as in the different access to information and cognitive resources (Macnaghten and Chilvers 2012; Stirling 2008; Wynne 2006). The relationship between the form of participation and the respective research is particularly relevant in deliberative participation formats where the medium of inclusion is intended to support factually correct, ‘reasonable’ research that serves the common good. After all, a lack of clarity about the role of participation can contribute to a disappointment among participants and lead to an additional burden for competent researchers, e.g. due to the increasing complexity of their work and communication (Rogers 2003).

2.1. Mini publics in technology research

The subproject ‘Doing Mini-Publics: Translocalisation of Politics’, funded by the Collaborative Research Centre ‘Re-Figuration of Spaces’ (CRC 1265 of the German Research Foundation DFG) from 2018-2021, has created a database for recording mini-publics from the 1970s to 2021 and mapped their distribution (Sayman and Hügel 2021). These and comparable databases such as the OECD ‘Database of Representative Deliberative Processes and Institutions’ (OECD 2021) show how mini-publics have become increasingly widespread worldwide, especially in recent years. The OECD database lists a total of 574 examples from around the world as of 2021 whereby 101 examples were added between January 2019 and October 2021 alone. It becomes apparent also by their documentation in social studies of science and technology assessment, that mini-public participation formats are increasingly researched and applied in upstream scientific and technological decision-making processes (cf. Kaplan et al. 2021).

Specific projects address e.g. the question of under what conditions and principles scientific research in the field of geoengineering should be carried out (Bellamy, Lezaun, and Palmer 2017) or what approach is preferred by the public for the US National Aeronautics and Space Administration to redirect asteroids into lunar orbit for study and exploitation (Tomblin et al. 2017). As the mini-publics database shows, over the past two decades, biotechnology or genetic engineering and neurotechnology or brain research have been among the most frequently covered topics within the ‘Science and Technology’ cluster (Sayman and Hügel 2021). Internationally, a citizens’ report was recently drawn up as part of the ‘Australian Citizen's Jury on Genome Editing’ in which randomly selected citizens of Australia developed recommendations on the conditions under which genome analysis technologies or the editing of the human genome could be acceptable. Notably, these recommendations were compared with the opinion of experts and the general public via survey and then translated into political recommendations for action (Nicol, Paxton, and Niemeyer 2022). In this context, a mini-public procedure in the USA should also be emphasised, which was not aimed at increasing the acceptability of genome editing but was openly framed regarding its outcome. Here it became clear that, from the point of view of the citizens involved, research on genome editing should be limited and deferred until healthcare access concerns are eased on a broader economic-political level (Nelson et al. 2024).

Mini-public participation has also become increasingly important in the field of AI in the health sector. In 2019, the UK's Greater Manchester Patient Safety Translational Research Centre (NIHR) carried out mini-publics in the form of two Citizens‘ Juries. In the context of the statutory data protection regulations, the question was whether to limit the complexity and thus the accuracy of AI so that its decisions can continue to be transparent and explainable to those affected by it. In the two Citizen's Juries (in Coventry and Manchester), 36 representatives of the population worked out their recommendations based on information over 5 days (NIHR 2019).

3. Planning cells for a citizens’ report

Planning cells for a citizens’ report as a new form of participation was developed in the 1970s by Peter C. Dienel with the aim of supplementing the expert-led political control apparatus with public policy-oriented decision-making processes based on a random selection of citizens. The method was developed at the same time, but independently of each other, when Ned Crosby in the USA introduced the Citizens‘ Jury as new deliberative participation procedure. Both forms of deliberative mini-publics with random selection of citizens, comparable to Consensus Conferences, Citizens ‘Assembly or Deliberative Polls, have spread worldwide since then (Dienel 1997, S. 168, 2002, S. 27).

Planning cells comprise a rather standardized participation procedures whereby groups of around 25 randomly selected participants are involved in a structured process with a content-based agenda on a socially relevant topic for 2–4 days. They receive information from experts, discuss in small groups with changing composition, and develop their recommendations. They receive an expense allowance for their participation and can be released from school or work obligations. The process and results are documented and published in a citizens’ report. As a rule, at least two or more planning cells are carried out parallel, but at different times, on a topic with the same procedure and the same information. The results from all planning cells are compared and merged into an overall result, whereby the individual results are also documented in the citizens’ report. The draft of the citizens’ report is first submitted to all participants, and only after their feedback is the final version of the citizens’ report prepared, handed over to the addressees or initiators of the participation process and published. The entire participation process is organized and moderated by a neutral process facilitator. Traditionally on place, planning cells for a citizens’ report have also been carried out online since 2020, driven by the requirements of the Covid-19 pandemic (nexus Institut 2023). Overall, the special features of the participation process are intended to enable an information-based, democratic, deliberative process and to derive recommendations that are oriented towards the common good (H.-L. Dienel and Böhm 2021; H.-L. Dienel and Franzl 2014; P. C. Dienel 2002, 2009).

3.1. Planning cells in technology research

A planning cell database of the nexus Institute has documented planning cells carried out primarily in Germany and other European countries since 1976 and comprises 77 entries (nexus Institut 2023). The method was initially used in Germany in the field of urban development. However, the participation procedure proofed particularly useful and suitable for the TA as well, because it can involve citizens in the development and evaluation of new technologies and services, raise previously unanswered questions, open up new perspectives and examine the social acceptance that should be incorporated into research (Dienel 1997, S. 163f, S. 175). Already in the first years after their introduction, planning cells for a citizen's report were used for participation of German citizens in technology-related topics, e.g. on new information technologies (1986), on ISDN (1990), or five years later on biotechnology or genetic engineering (nexus Institut 2023).

The questions addressed in these examples each aim for a framework for how existing technologies, including their inherent advantages and disadvantages, should be used. However, planning cells have also been tested in technology-related research projects to develop technologies in such a way that their design approach supports social values. For example, in 2017, the BMBF-funded research project ‘Innovative Transport Access Systems through User Participation’ (nexus Institute 2017) carried out planning cells to derive recommendations of the citizens involved for the development of ticket systems based on Radio Frequency Identification (RFID) technology. The research network ‘Research campus Mobility2Grid’ deals with the connection between the energy grid and the electrification of transport and has published a citizens’ report on technical links for ‘Intelligent Energy and Transport Transition in Berlin City Districts’ (Research campus Mobility2Grid 2014).

4. Participation in the VI-screen project

To illustrate the practical use of planning cells for a citizen's report in the research-based development of new technologies, the implementation and results of the participation procedure in the context of VI-Screen are presented below. All steps, contents and results are documented in detail in the citizens’ report (TU Berlin, FG ARTE 2023). As mentioned at the beginning, against the background of the Covid-19 crisis (2020–2023), the project aimed to develop a monitoring system that would contactlessly measure the characteristics of infectious respiratory diseases. In the future, such a system should be able to detect a risk of infection in individuals in order to contribute to the protection of critical infrastructures such as hospitals and care facilities.

Participation in the research-based development of the VI-screen system had the following objectives: Firstly, to give citizens and clinicians the opportunity to gain insight into the research at a very early stage and to contribute their perspective to the further development of the technology through knowledge – and discussion-based recommendations. Secondly, to involve the researchers in the formulation of sub-topics and questions for the participation process, in the provision of information in the planning cells and to use the participation results for further technology development. The latter referred in particular to the ethical and legal framework conditions for data processing by the VI-Screen system. In the following, the implementation of the planning cells for the citizen report in the VI-Screen project is explained in order to discuss to what extent these two objectives have been achieved.

The Department of Work, Technology and Participation of the TUB in the research project delivered two planning cells, offset by one hour and two days each, whereby a random selection was to be drawn from the population in Homburg/Saarland (i.e. the location of the project partners UKS and UdS) and from employees of the UKS in order to involve a total of 50 people. Preparation and implementation or process support were carried out by the Department of Work, Technology and Participation of TUB, which was not involved in the technical development of the monitoring system. The team of the department also includes the authors of this article.

The process support mainly comprised the following tasks: (1) Accompanying development of the agenda contents, their sub-topics and questions for the small groups with the project consortium, i.e. with the scientists who are developing the VI-Screen Monitoring System, desk research and acquisition of expert speakers for information transfer in the procedure; (2) Implementation of all necessary organizational steps in the participation process, e.g. application for the drawing of the random sample and sending of the invitations with written, lay-friendly information about the participation process and topic, return management, organization of rooms, work materials, catering; (3) Implementation of the two-day planning cells, i.e. moderation and event management as well as photo documentation of all results, writing the first draft of the citizens’ report on the basis of the documentation, sending the version to all participants, incorporating all feedback from participants, creating the final version of the citizens’ report, organizing the handover event; (4) Communication of the results to the consortium and review of the extent to which the results of the participation were taken into account in a new version of the document ‘Framework conditions for data processing’ of the system.

The planning cells were implemented after a technical experimental set-up of the VI-screen monitoring system had been implemented and the collection of the raw data from test subjects had already begun. The AI-based evaluation of the data of infected or uninfected persons was not yet started at this time, so that it remained unclear to the participants of the planning cells whether the system would actually be able to detect respiratory infections contactlessly in the future.

The planning cells took place from 22/09–23/09/2022 from 09:00 am to 06:00 pm. For better comprehensibility and motivation to accept the invitation, the following title was chosen for the planning cells and the citizen's report: ‘Citizen's report on contactless infection control at the clinic door – participation in the development of new technologies in the VI-Screen research project’.

The procedure was divided into 8 work units of 1.5 hours each. In each work unit, information was first provided by experts through lectures with presentations, and participants were able to ask questions. The information transfer section lasted around 30 minutes each. Afterwards, the participants discussed in 5 unmoderated small groups using guiding questions for 30 minutes each. The composition of the small groups changed after each work unit. Afterwards, the results of the small groups were presented in the plenary session and clustered by the participants. Afterwards, all participants received a predetermined number of points with which they could individually evaluate the result clusters. The overall weighting of the recommendations, as documented in the citizens’ report, was later determined by counting the points awarded.

With this, the procedure for the preparation and implementation of the planning cells largely corresponded to the above-described main features of the planning cells procedure for a citizen's report. Deviations in the procedure primarily concerned the selection of participants. While a random selection from the population of Homburg was possible via the population registration office, the clinic employees could not be invited by a random draw due to administrative requirements but were addressed via the internal newsletter of the UKS.

Although the registration numbers were initially well above the target number of 50 people, due to a re-emerging wave of the Covid-19 pandemic, there were an unusually large number of cancellations very shortly before the start of the planning cells, which could no longer be filled by people on a waiting list. Overall, a group of 41 people, which is very heterogeneous in terms of gender, age, place of residence in Homburg, immigration history, education and occupation, but not representative of the population, participated in the planning cells.

The agenda for the planning cells, i.e. the formulation of the main topic or the sub-themes, was primarily based on the topics of the ethics application for the VI-Screen project. This ethics application had to be made at the beginning of the research project for research on humans or data collection with patients, legally reviewed and also approved by the responsible ethics committee. An essential part of the ethics application was the accompanying document ‘Framework Conditions’, which explains the measures taken to protect personal data, to prevent data misuse and to comply with ethical principles in the development of the AI-supported VI-screen system. The document Framework conditions initially included the measures taken by the research team during data collection and processing in the ongoing VI-Screen project. The recommendations from the citizens’ report should be used in particular for the development of scenarios that provide for a further development of the VI-screen system with a trial regular operation with accompanying research and a potentially autonomously running operation of the VI-screen system. The participatory document Framework conditions should thus support the researchers in the development of the VI-screen system up to its possible use in practice.

The selection of sub-topics and questions therefore focused on the ethical-legal aspects with a focus on the framework conditions of data processing for the VI-screen system. In addition, consideration should also be given to whether advantages are seen in the use of such a system, which concerns must be taken into account so that a potential use of the system could be accepted, and for which areas of life the autonomous use of such a system would be considered ethically justifiable. The key questions that were to be discussed by the small groups in the work units were formulated by TUB in consultation with the consortium partners. Sub-topics, key issues for the discussions and main recommendations are summarized on the following in Table 1.

Table 1. Topics, key questions and recommendations of the participants from work units 1–4 of the Citizens’ Opinion on VI-Screen.

Work unit (WP) or sub-topic	Key questions for the small group discussion	Main recommendations (keywords based on the summary of recommendations from (TU Berlin, FG ARTE 2023))
WP 1: Welcome, introduction to the method and procedure, presentation of the ‘VI-Screen’ project	Work unit was shortened in favor of more time for AP 2, there was no small group discussion here yet.	Work unit was shortened in favor of more time for AP 2, there was no small group discussion here yet.
WP 2: Presentation of the planned virus screen system	From your point of view, what advantages does the presented virus screen system have, e.g. compared to anti-gene rapid tests or other test systems known to you or for safety in a clinic?	Possible advantages of the planned Virus screen system: created as a system capable of learning can detect various virus-related respiratory infections in a standardized manner environmentally friendly and sustainable, does not require any consumables and thus avoids waste contactless, non-invasive measurement method that no longer requires a nasopharyngeal swab possible time and cost savings
WP 3: Data protection requirements in the VIRUS screen system	The VIRUS screen system would store video footage of faces for only a few minutes: What concerns might you still have? And what could help you to stop having these concerns?	Main concerns about the virus screen system: Security of personal data, such as video recordings of faces, could not be guaranteed manipulation and misuse Main recommendations for reducing concerns: superordinate supervisory body for regular monitoring of compliance with all data protection measures the greatest possible transparency about the use and backup or deletion of data for persons measured by the system and for the public Avoid storing data, sufficiently informed consent to measurement and short-term storage lasting a few minutes after measurement, proof of deletion of overall image recording and personal data
WP 4: Risks to critical infrastructures due to cybercrime	What are your biggest concerns about cybercrime when you think of the Virus screen system presented?	Main concerns and recommendations regarding cybercrime: Misuse of data, identity theft, manipulation of data to falsify measurement results Impairment of system functions Theft of the system and extortion through external attacks Every effort must be made to avoid risks from cybercrime
WP 5: Philosophical-ethical questions when using screening systems in(clinical)everyday life	In which fields of practice can you imagine the use of screening systems from an ethical point of view? In which fields of practice would you reject the use of screening systems from an ethical point of view?	Ethically justifiable fields of practice: Healthcare and transport and travel Field of education less supportive Ethically justifiable fields of practice: Private and Leisure Facilities of daily life
WP 6: Artificial intelligence in everyday work – A look from the employee perspective	As a clinician or if you put yourself in the perspective of clinicians: What should be considered when introducing new technologies with artificial intelligence, such as the Virus screen system?	Recommendations from an employee's perspective on the use of artificial intelligence in everyday work: Employees must be informed and involved as early as possible, they must learn what, how and with what consequences the use of the system is planned and which training courses are planned Capture and address concerns and objections in a timely manner The introduction of AI should not lead to the reduction of personnel, it should only be used for relief, support and assistance, people must retain control and decision-making authority
WP 7: Image use and image processing in research	If a video of you were used for research, what would be most important to you so that you would consent and maintain that consent? If the virus screen system were to store pseudonymised information on symptoms for further improvement of the system, under what conditions would you agree to this?	Recommendations for image use and image editing in research: Image data should only be used for the purpose stated in the research Purpose should serve the general public It should be comprehensively informed about the use and protection of the data and this should be guaranteed If the data is used for another research purpose, there must be renewed consent
WP 8: Summary of results, evaluation questionnaire, organizational and closure	In this last work unit, according to the procedure of the participation procedure, no small group discussion took place.	In this last work unit, according to the procedure of the participation procedure, no small group discussion took place.

In two of eight work units, the information was provided by a scientist from the project consortium who was significantly involved in the technical development of the system. This also included the presentation of the experimental setup of the planned VI-Screen system, including the possibility for participants to have themselves measured on a test basis. All other information was conveyed by external speakers not associated with the project with corresponding technical expertise. Detailed information about the process as well as all results from the two planning cells and all individual results can be found in the published citizen report (TU Berlin, FG ARTE 2023).

In addition to the motivation to participate and the evaluation of the implementation of the planning cells, various statistical data on the description of the composition of the group as well as the attitude towards technology and the Covid-19 pandemic were recorded in an anonymous questionnaire, which all participants completed at the end of their participation in the planning cells. As provided for in the participation procedure, the process and the results of the planning cells were then documented in a citizen report agreed with the participants.

After the approval of the citizens’ report by the participants, the consortium was informed in advance of the results. The official handover of the citizens ‘ report to the project consortium took place six months after the planning cells in the UKS premises in Homburg. After the citizen's opinion was available, the recommendations were incorporated into the ‘Framework Conditions’ document as far as they can be considered and used for the further development of the VI-screen system.

5. Process evaluation and discussion

Against the background of the previous presentation of the use of planning cells for a citizens’ report in the VI-Screen research project, the achievement of the participation goals is evaluated and discussed below.

5.1. Citizens-related participation goals

Citizens and clinic personnel should be given the opportunity to gain a very early insight into the research and to contribute their perspective to the further development of the technology through knowledge – and discussion-based recommendations. The prerequisites for this were created mainly by the implementation of the planning cells in accordance with the standard procedural requirements of the participation procedure as mentioned in chapter 3. However, the following aspects should be discussed critically:

Early participation in the event of remaining uncertainty about technical feasibility: The participation of potential users or affected persons at the earliest possible stage is becoming increasingly important, especially in view of technical path dependencies and widespread social fears about AI applications in the health sector (see below), which also affect a system such as the VI-screen system. In the context of the parliamentary TA, Grunwald calls for:

Stakeholders, civil society organizations and citizens or patients and those affected must not only be involved in the market launch but must also be involved in the development and preparatory steps of a technology launch. Although this does not guarantee a constructive course of risk debates, it is likely to be a necessary condition for this in an open and pluralistic society (Grunwald 2022, S. 557)

In the present case, the technical innovation was still at a very early stage of development. The question of whether reliable prediction of virus-related respiratory disease is possible by means of AI-based evaluation of the sensor data obtained had not yet been answered at the time of participation. Those addressed by random selection therefore did not have to see themselves directly affected by the research. In this context, the ‘participation paradox’ should be mentioned (Reinert 2003) as a theoretical influence on the willingness to participate. Accordingly, the more concrete a project and its influence on one's own lifestyle already is, the greater the interest in participation in planning processes. However, the participatory leeway decreases with increasing planning maturity (Hirschner 2017). The authors also see a transferability of this concept in research-based technology development. Due to the freedom from scientific experiments for the ‘production of errors, mistakes and failures’, the participation paradox (Krohn and Weyer 1989) can even represent a hurdle to early participation to a particular extent or express itself in a lower willingness to participate.

Even in Project VI Screen, the originally targeted number of 50 people for the planning cells was not reached. However, most cancellations were due to illness or illness of colleagues due to the Covid-19 pandemic, so that no empirical evidence can be found on the influence of the participation paradox.

Heterogeneity of those involved: Potentials of AI applications have already been confirmed for research, diagnosis, therapy and care in healthcare (Rasche, Reinecke, and Margaria 2022, S. 54f.). However, these potentials are faced with various legal, socio-political and economic hurdles. The lack of acceptance from users is one of the biggest hurdles:

Especially in medicine, AI applications are associated with high expectations and at the same time fears […]. The discussion about the limits of what is ethically desirable and meaningful is in full swing and is fueled further with each new AI application. (ibid. S. 60)

In planning cells for a citizens’ opinion, the heterogeneity of participants in the small groups is a prerequisite for diversity of perspective in their discussions and the deliberative, argument-exchanging process of developing recommendations. It is also an important prerequisite for the usefulness of the recommendations for researchers in order to improve the ability of developments to connect to social value relationships (Dienel 2002). In the context of technical innovation, this can support potential diffusion and subsequent application of the technology.

In relation to the group of ultimately 41 participants, an attempt was therefore made to achieve a heterogeneous composition. This was achieved in terms of demographic characteristics such as gender, age, place of residence or district, immigration history, education and occupation and can also be substantiated by the consistency of main recommendations of the citizens’ opinion with results of representative surveys on attitudes towards AI in general and in the healthcare sector in particular. Thus, the representative survey of the ‘Consumer Study 2021 – Safety and Artificial Intelligence’ confirms fears and skepticism of the population in the field of application of AI in medical diagnostics (TÜV-Verband e.V. 2021, S. 13).

Nevertheless, it should be noted critically that fewer clinic employees than intended were involved in the application example. It was not granted by the clinic, despite information from the staff council, for data protection reasons, to draw a random sample from the population of clinic employees. Here, it would be desirable to make it possible to draw randomly from institutions in the future. On the other hand, it is critical to note that people who are generally not or very little informed about or interested in new technologies and people who have an indifferent attitude towards the Covid-19 pandemic were not represented among those involved. Here, it should be considered what, in addition to the elements already provided for in the participation procedure, which are intended to facilitate participation (motivational invitation, low-threshold registration option and contact persons in advance, reimbursement of expenses, possibility of recognizing participation as further education), could increase the interest in participation of persons with a negative or indifferent attitude in order to enable an even more diverse and controversial discourse.

Agenda setting: Topics and decision-making areas of participation are defined in traditional top-down approaches by the responsible project managers. Due to these differences in power and within the framework of formalized structures, the participants are then no longer able to put the questions relevant to them on the agenda, which cannot least lead to a disappointment of the participation euphoria (Abels and Bora 2004) among the participants, to frustration and a boycott of the participation process by the participants (Rogers 2003). With regard to the specification of tasks in planning cells, Peter C. Dienel also notes that the participants ‘existing awareness of the problem is not exhausted in this way. In addition, the topic specification ‘has the disadvantage that the planning cell has to do with its work without the possibility of identification associated with the self-selection of a topic and the additional motivations given with it’ (C. Dienel 2002). However, he considered the disadvantages of free choice of topic to be ‘more serious’, and this led to excessive demands and unreasonable time for the participants as well as a more difficult integration into the overall planning and preparation of the information input (ibid.).

Further in project VI-Screen, the possibility of the knowledge – and discussion-based development of recommendations by the participants was limited to the sub-topics and questions that were included in the agenda for the planning cells. Participants’ satisfaction with these aspects was recorded as part of the anonymous evaluation questionnaire they completed at the end of participation. This shows an on average very positive evaluation with regard to invitation, procedure or program, moderation and event organization (TU Berlin, FG ARTE 2023). However, the preliminary information and the questions and discussions in the small groups were assessed as less clearly satisfied (ibid.).

One reason may be the lack of involvement in the agenda setting process of VI-Screen. As shown, the development of the questions for the small groups was carried out by TUB in coordination with the consortium of the research project (ibid., chap. 7.2). As the overview of the questions in Table 1 shows, some complex questions were also given to the small groups for discussion. These corresponded to the content interests and questions of the researchers. Even if their formulation was accompanied by TUB as partners not involved in the technological development, who prepared and carried out the planning cells: there was no external person involved in the development of the sub-themes and questions. There was no explicit provision for influencing the fundamental question of whether the system should be (further) developed. Here, it is necessary to consider how the process of agenda setting could be adapted to take greater account of the external or stakeholder perspective, but at the same time to address the disadvantages P. C. Dienel (2002) mentioned. This aspect will be taken up again below.

5.2. Researcher-related participation goals

Researchers were supposed to participate in the formulation of sub-topics and questions for the planning cells, to be included in the provision of information in the process and to be empowered to use the participation results for further technology development. The first objective was ensured by the joint development of the agenda in the consortium. The second objective, the discussion of the results of the participation, was at least structurally secured by the inclusion of recommendations from the citizens’ report in the framework conditions for a potentially autonomously running VI-screen, insofar as this appears to be possible in terms of content and technology. The latter point is particularly positive for the application example in project VI-Screen. However, the following aspects should also be discussed critically:

Requirements for transparency and dissemination of information: On the one hand, it is to be regarded as positive that participation in communication virtually entails a compulsion to concretize and formulate research in a way that is understandable to layperson. Only in this way can transparency be achieved through research, provided that individual components, such as the algorithms for technological development, do not have to remain confidential. Tennis Doe has summed up the downside for participatory health research as follows: ‘It will be harder to do your research, take longer, cost more money and be significantly more complicated when making even basic decisions.’ (Doe 2000, S. 54 cited after Flieger 2009, S. 7)

In the application example discussed here, apart from the partners preparing and carrying out the participation procedure, the researchers had no previous experience with deliberative participation formats. In part, although not in the concrete application example, skepticism of the researchers must be overcome that participation creates concrete added value for participants and researchers (Rogers 2003). In addition, this results in requirements for the preparation and support of researchers for participation.

As already stated elsewhere (comp. Dienel and Böhm 2021, S. 19f.), this concerns measures for greater openness of researchers to a greater delimitation of their role and the role of those involved, as well as higher requirements for transparency, easy-to-understand communication of information and willingness to dialogue, communication and reflection. Concrete adjustments to the participation procedure in this regard are taken up again in the recommendations.

Accountability for the consideration of recommendations: Even if citizens expect their perspective to be taken into account more through greater dissemination, better anchoring and greater support for participation in research (BMBF 2023a): there is no obligation to take into account the results developed in a participatory manner.

This also applies to the planning cells for a citizens’ report in project VI-Screen. The results were only of an advisory nature and the researchers’ approach is protected from ‘freedom of research’ within the framework of good scientific practice. In the event that recommendations are not taken into account, there is a maximum of one, at least socially informal and democracy-theoretical, accountability or feedback request from the researchers to the participants. Over the course of several days, they dealt with a topic that was foreign to them, not least for the researchers or their shared research interest (Abels and Bora 2004).

This aspect of bindingness in dealing with the results of participation has an impact on the relationship between researchers and participants. Without a mandatory consideration of recommendations, an appreciative and cooperative attitude of the researchers will inevitably become central to the ability of participants to help shape the research process. However, research and knowledge generation should be seen ‘not as a privilege of scientists, but as a co-production of all researchers involved’ (comp. Hartung, Müller, and Tosun 2020, S. 2). We also take up this aspect again in our recommendations (see below).

Challenges of layperson's comprehensibility and complexity: In the context of participation in research, scientists and external experts must convey complex knowledge to layperson. It is necessary to weigh up the complexity and depth of content on the one hand and the necessary comprehensibility and inclusiveness on the other. A good ability to communicate knowledge or science in a way that is understandable to layperson is therefore essential both for the participants and for the quality of the participation results.

In the application example of VI-Screen, this concerned in particular the topics of AI, machine learning, image processing and ethics. This presents a dilemma: the more comprehensible and therefore less complex the key questions for the small group discussion are selected and formulated and the knowledge is imparted, the more general and thus possibly also less complex and connectable the recommendations from the participation could be. It should also be critically discussed to what extent the participation recommendations can be connected to the researchers’ work at all. On the one hand, this concerns the question of whether the quality of the response for the researchers was high enough to ensure that the recommendations could be incorporated into further research work. On the other hand, this concerns the question of to what extent the selection of the questions was expedient for the researchers’ interest in knowledge. While individual sub-topics, such as requirements for data protection and security against cybercrime or image use and image processing in research, were directly related to the framework conditions of the VI-screen system, other sub-topics, such as ethically justifiable areas of application, were aspects that the researchers did not directly influence.

The above-mentioned tendency of main recommendations from the citizens’ opinion to agree with representative surveys allows positive conclusions with regard to the heterogeneity of the group of participants but could also be viewed critically if the recommendations are not more differentiated. Although the citizens’ report actually offers more differentiation than is reflected in the summaries of the main recommendations, there are limits to the complexity and information conveyance, which are also related to the fact that the tasks in planning cells must meet the criterion of ‘workability’: ‘Tasks assigned to a planning cell must be tailored in such a way that they can be mastered in the time available. The prerequisite for this type of topic assignment is the ability to limit a problem’ (C. Dienel 2002). A practicable middle ground must be found, which includes, in particular, the preparation and support of researchers in terms of knowledge and science communication as well as the drafting of the research agenda, which we will also briefly take up below.

One-time participation in the research process: One ideal of participatory research is that it should be realized ‘in different phases of project implementation’ (Farin-Glattacker et al. 2014, S. 4) with varying intensity at the stages from consultation to participation (ibid. S. 6). It is questionable how this aspiration can be met in the context of often time-limited research projects for the development of new, AI-based technologies.

In the VI-Screen research project as well, participation was only realised at the beginning of the research project. One possible limitation of the participation format carried out in the application example is therefore that the planning cells in the process represent a rather isolated procedure, which is completed over a period of a few days and the results of which are recorded in a citizen's report. Assuming the technical success of the company, it can be assumed that further interests will influence the development until a possible exploitation of the VI-screen system. In this context, it should be asked whether a citizens’ report in its classic appearance is the appropriate vehicle to transport results and make them comprehensible to researchers (even later). The citizens’ report provides participants, researchers and the public with transparency and information about the content, process and results of participation. However, it is also possible that, in parallel with this and initially due to the protection of technological developments, an alternative approach is needed that supports researchers in their further developments.

Due to the framework conditions formulated internally in the research project VI-Screen, this case was possible in the application example and was assessed as useful by the researchers. Such an approach could also be useful for other applications.

6. Recommendations and outlook

The mini-public participation procedure planning cell for a citizens’ report in project VI-Screen was used to give citizens and clinicians the opportunity to gain a very early insight into the research and to contribute their perspective to the further development of the technology through knowledge – and discussion-based recommendations. Furthermore, the researchers should be involved in the formulation of sub-topics and questions for the participation procedure and the provision of information in the procedure and enabled to use the participation results for further technology development.

In summary, these objectives were met, whereby the authors were also able to derive adaptation potentials from the reflection of the applied method and the discussion of the developed recommendations. To anticipate here, taking these recommendations into account would mean more resources for the use of the procedure. From our point of view, however, it would also increase the chances of greater willingness to participate, broader heterogeneity, participation in agenda setting, more intensive preparation of researchers, better response quality and connectivity, as well as ongoing participation in the research process. From our point of view, however, it would also increase the chances of greater willingness to participate, broader heterogeneity, participation in agenda setting, more intensive preparation of researchers, better response quality and connectivity, as well as ongoing participation in the research process (C. Dienel 1997).

1. Early participation and broader heterogeneity through approachable approach and more fees: With Strothmann (Strothmann 2020), we recommend intensifying the invitation process by addressing people personally who do not initially register after receiving the invitation. Strothmann refers to this kind of personal approach as ‘searching lottery procedures’ and has used them several times in two municipalities (ibid.). As Strothmann emphasizes, this approach involves additional effort, but is intended to increase the chance of reaching those who are classically more difficult to reach for participation. These include, for example, people with lower educational qualifications and lower social resources or with an indifferent to negative attitude towards citizen participation or the topic that will be dealt with in a participation process (ibid.). In the example of the planning cells for the citizens’ report in the VI-Screen research project discussed here, people with lower educational qualifications and incomes were also among the participants. However, as mentioned above, people with a negative or disinterested attitude towards new technologies and the Covid-19 pandemic were missing. Since it will always be a voluntary participation and all legal requirements of data protection must be observed, there are limits to the approach to be taken. However, as Strothmann describes (ibid.), it is possible to send another letter to those who do not register with a request for a personal, telephone or digital conversation and to point out that there is a special interest in reaching, for example, people with an indifferent to negative attitude towards citizen participation or the topic. Although this approach requires a greater willingness to deal with conflicts, it can increase the diversity of opinions in the planning cells. A further recommendation here, even if the participants in the planning cells for the VI-Screen research project were satisfied with the amount of the expense allowance, of 50.00 euros per day, according to the evaluation questionnaire, was to increase the remuneration as an incentive and appreciation in the future. Peter C. Dienel originally provided for a loss of earnings for the planning cells, i.e. depending on income, people received more or less money for their participation. This seems rather helpful, for example, to reach more self-employed people and people with very high incomes and has not prevailed as a procedure. Instead, we propose a higher overall reward for participation for all participants.

2. Enable citizen participation in agenda setting through a two-stage process: For large infrastructure projects, it is already proposed to offer citizens, stakeholders and associations the opportunity to participate already in the phase of designing a participation process. This procedure is referred to as ‘participation scoping’ and is described as a method ‘⁣ ⁣with which informal participation processes can be prepared. The project initiator or project developer designs the content, formats and timetable of the participation precisely in advance and then coordinates them with the participants⁣⁣.’(Bertelsmann Stiftung 2017, S.3).⁣ Based on this approach and on the ‘round table’, which is ideally intended for the use of planning cells in the political sector for agenda setting (Gesellschaft für Bürgergutachten 2015, S. 2), we propose a two-stage process for the use of planning cells for citizens’ reports in research, which should enable the participation of randomly selected persons already during agenda setting. In a first phase, the agenda, including sub-topics and key questions, would be drafted for the small group discussion with the researchers, as was also the case with the planning cells for the citizen's report in the VI-Screen research project. In a second phase, this agenda would be presented to randomly selected citizens, and they would be given the opportunity to make recommendations for adapting or supplementing sub-themes and key questions. Due to their bias, these citizens would not participate in the planning cells themselves and would be obliged to remain silent about the agenda until the planning cells were carried out. This approach would have the following advantages: even earlier participation, greater comprehensibility of the key questions and more consideration of aspects that are relevant to the research topic not only from the perspective of the researchers, but also from the perspective of citizens. If it is informed that citizens have already participated in the agenda setting, this could also contribute to increasing the willingness to participate and more heterogeneity of the stakeholder groups, in addition to the procedure already listed above.

3. Offer participation coaching to prepare researchers and to increase response quality and connectivity: Researchers are experienced and experienced in reporting on their work in qualification work, articles in journals and contributions to conferences. They are expected to communicate with professional expertise, disciplinary terms and content complexity. Scientific institutions usually have a press department that informs about the work of researchers, even in layperson's terms, in a concise form. In the field of science journalism, research is communicated in more detail and usually with its own disciplinary qualification. Science communication in planning cells, i.e. information about the research process and content, requires researchers to communicate in a sufficiently complex and professional and at the same time easy-to-understand manner. When using planning cells for a citizen's report in research, there should be an offer to prepare for this type of communication. Participation coaching, which we recommend here for researchers in preparation for their role and task in planning cells, would, in addition to science communication, also impart knowledge about participation and an open attitude, which is necessary to appreciate, seriously discuss and, as far as possible, consider processes and results of participation. This also includes a discussion of possible reservations or skepticism about participation. Another part of participation coaching should support researchers in defining sub-topics and areas in which their research actually still has scope for taking into account recommendations from participation. Equally important is the examination of the degree of complexity that is required for participation in the transfer of knowledge, the derivation of the sub-topics and the formulation of the key questions, so that the quality of the answers or the quality and specificity of the recommendations developed by the participants in the small group discussions is connectable and thus receptive to further research.

4. Obliging researchers to account for the consideration of participation results: Planning cells for citizens’ reports are a process of informal participation, i.e. participation that is not regulated by law. Thus, there is no obligation on the part of the initiators or at the same time the addressees of the participation procedure to take the recommendations into account. However, it should be assumed that there is a serious interest in the recommendations and at least feedback should be given to stakeholders on how their recommendations have been included or for what reasons they have not been included. This part of the entire participation process, which refers to the time after the planning cells have been carried out and the citizen's report has been published, is often neglected. Here we recommend a greater commitment, which should already be included in the research proposal and thus also in the research funding. Researchers would then undertake to report on the consideration and justified disregard of recommendations at least in a document that is internal to them, but which must also be accessible confidentially to the research-promoting institution. Documented accountability and reflection could also benefit the methodological further development of the participation format in research. The use of planning cells for the further development of the framework conditions of the monitoring system in Project VI Screen was a suitable means here.

5. Provide for ongoing participation with more interaction between researchers and participants: Finally, the importance of participation in the overall temporal context of a research project that aims at the development of a market-ready product in the long term should also be considered. If the participation procedure Planning Cells for a Citizens’ Report is used for very early participation, a continuous continuation in a form that allows more interaction between researchers and participants for a joint search for the best solutions is conceivable. In the further course, the researchers give feedback to the participants on the recommendations that they were able to include, explain why they could not include certain recommendations and discuss with the participants what alternative solutions might still be available. This approach leads to a new constellation of roles between participants and researchers, in which they interact more on an equal footing. This ongoing process is not linear, but rather circular in nature, allowing for a greater shift in information, interaction and feedback, and can go as far as follow-up applications for further research, product maturity and business plans. In summary, it is essential that researchers and participants interact in such a way that all the potential of participation for further developments and new ideas in the process can be optimally used and the acceptance of the participants for the solutions developed can be increased.

With regard to these recommendations, it should be noted that the present case study is an evaluation by the process support carrying out the participation procedure, but not an independent scientific evaluation by external scientists (Ayano 2021). The recommendations are therefore only exploratory in nature. Their empirically reliable confirmation or criticism in the context of their testing is desirable. In summary, however, the authors suggests that planning cells for a citizen's report are a suitable method for participation with random selection in technology research, considering some methodological recommendations. Our suggestions may mark a contribution to this means. Especially in the context of a rapidly accelerating digitization of our living and working environments through the use of artificial intelligence, we see an increasing relevance to strengthen meaningful applications of participatory formats in research-based technology development.

Disclosure statement

No potential conflict of interest was reported by the author(s).