The future of manufacturing: Utopia or dystopia?

Digital manufacturing technologies (DMTs) have the potential to transform industry productivity, but their introduction into the workplace is often a complex process, requiring not only technical expertise but also an awareness of ethical and societal challenges surrounding human–system integration. Concerns about the introduction of new technology have been prevalent throughout history, and exploring public perceptions of these technologies can provide insight to help address such cultural anxieties. However, evaluating user perceptions of futuristic technology is difficult, requiring novel approaches to provide context and understanding. To explore users' perceptions of future DMTs, we applied the ContraVision technique in a questionnaire‐based study. Participants viewed films, representing fictionalized utopic and dystopic visions of what the future of these DMTs might involve, and a questionnaire probed the perceptions of the technologies afterward. Findings showed that irrespective of the way technology was portrayed, participants had concerns about the ethical and responsible implementation of these tools. Participant responses were analyzed to identify key challenges for policy surrounding DMT implementation in the future of manufacturing.


| INTRODUCTION
In a concept known as Industry 4.0, advances in digital technology are creating new opportunities to improve productivity within the manufacturing industry (Pereira & Romero, 2017). Digital manufacturing technologies (DMTs) found within Industry 4.0 utilize information and communications technologies to support datadriven decision-making and performance. These include systems such as industrial Internet of Things (IoT) technologies (Asplund & Nadjm-Tehrani, 2016;Sisinni et al., 2018), automation and robotics (Javaid et al., 2021), virtual and augmented reality (Ariansyah et al., 2022;Erkoyuncu et al., 2017), and human physiological sensing (Argyle et al., 2021;Ariansyah et al., 2020). Industry 4.0 has the potential to transform manufacturing; however, as found in historical accounts of eras involving major technological change, effective integration relies on a design paradigm that considers the humancentered impacts of the new systems (Autor, 2015). Although DMTs have many potential benefits toward improving organizational productivity, system flexibility, increased product quality, and reduced environmental impact, it is recognized that significant challenges surround their implementation, for example, relating to addressing worker and societal concerns, the requirement for highly trained users, and cyber security risks (Sony, 2020). Research is needed to identify societal concerns involving the risks created by such systems, and neglecting this may limit successful technology implementation programs and lead to the failure of sound technological solutions (Morgan, 1997). In this study, we investigate public perceptions and attitudes toward DMTs through the demonstration of a novel design fiction method, the ContraVision technique (Mancini et al., 2010), with a view toward identifying challenges and opportunities for the design of future manufacturing systems.
In the present state of manufacturing, considerable research has focused on identifying barriers and enablers to Industry 4.0 adoption.
At the organizational level, barriers to Industry 4.0 readiness include legislative limitations or lack of standards, management issues, and a lack of workers with knowledge or experience with Industry 4.0 technologies (Stentoft et al., 2019). Successful implementation of DMTs such as distributed data technologies requires internal management of such systems to be at low cost and at a scale manageable with organizational resources (Maple, 2017). Challenges also exist around developing standardized protocols for technology implementation, for example, to ensure the security of IoT data in mobile versus nonmobile systems (Maple, 2017). At the individual level, perceptions and attitudes toward DMTs play a significant role; perceptions of artificial intelligence (AI) technologies are a significant factor in the adoption of such technologies (Bitkina et al., 2020). In terms of enablers of DMTs, communications on potential impacts of new technology may benefit from a user-centered approach, customized to target audiences (Morris et al., 2005), and focusing on specific applications rather than general use (Castell et al., 2014). Castell et al. (2014) surveyed the British public on attitudes toward science and technology, with a particular focus on the use of robotics across several domains. While 89% of the respondents stated that they had some familiarity with the use of robotics in manufacturing, the authors identified an effect of context on attitudes toward a technology; findings indicated that individuals who had learned about robots in positive contexts, such as safetycritical applications, had higher levels of positivity toward the technology.
Understanding public perceptions is important because it is a critical early-stage step toward effective integration of technology into workplaces such that this integration occurs in an ethical and sustainable manner that does no harm to workers in the system. A good understanding of public perceptions surrounding an issue can help to support effective education, communication, and policy formation related to the issue (Morgan, 1997). However, a challenge exists in that it is difficult to assess attitudes toward something that respondents do not have direct experience with, for example, with systems that are in development or not easily accessible. During the early phases of design, it is common for user-centered design methods to inform product or system development. In the human factors and ergonomics (HFE) community, methods often include interviews with stakeholders to derive specifications and requirements (Maguire, 2001), usability evaluation of prototypes (Abras et al., 2004;Argyle et al., 2017), and observational methods such as contextual inquiry to identify how users interact with the technology (Beyer & Holtzblatt, 1999;Raven & Flanders, 1996).
Nevertheless, user-centered design methods are often limited when it comes to assessing futuristic technology that may not be sufficiently developed to obtain robust user feedback. In response to this limitation, Mancini et al. (2010) proposed the ContraVision technique, a method for eliciting users' reactions through the presentation of opposing viewpoints in fictional narratives to potential end users.
Previous research has used the ContraVision technique, a method grounded in the human-computer interaction (HCI) domain, to identify potential user attitudes toward specific futuristic technologies, including those that do not yet exist or are not fully deployed (Mancini et al., 2010). The core of the technique centers around the creation of two opposing narratives: one, a "highly positive utopic" vision of the technology, and the other, a "highly negative dystopic" vision. Narratives are intended to engage the participant, and these can be told via film, illustration, audio recording, or written text (Mancini et al., 2010). In its first application, Mancini et al. (2010) used the ContraVision technique to explore users' reactions to wearable technology for assisting with health care and well-being. The authors concluded that presenting the same technology from two contrasting points of view elicited a wider spectrum of responses from the participants than by only presenting one viewpoint. The technique has been used in similar applications, for example, to explore users' understanding of and engagement with future smart grid technologies (Goulden et al., 2014), requirements for engineering adaptive software for IoT systems (Bennaceur et al., 2016), and societal acceptance of domestic energy Demand Response programs (Naghiyev et al., 2022).
From a methodological perspective, we argue that the Contra-Vision technique offers a useful approach for exploring future systems design in a way that has previously been underutilized in HFE research. The development of methods for designing and analyzing future systems from a human-centered viewpoint has often been within the purview of HFE, most notably from the systems ergonomics perspective (Wilson, 2014). In addition to user-centered design, methods employed in systems ergonomics offer a balance between qualitative and quantitative insights, integrated throughout the system design life cycle. ContraVision complements traditional HFE design and evaluation techniques, for example, cognitive walkthroughs (Mahatody et al., 2010), task analysis methods (Crandall et al., 2006), and Cognitive Work Analysis (Read et al., 2015;Salmon et al., 2016), among others. There has historically been close alignment between HFE and HCI methods, with approaches such as scenario-based design (Carroll, 1997), design fictions (Brown et al., 2016;Grand & Wiedmer, 2010), participatory design (Rogers et al., 2022), and ideation cards (Lockton et al., 2010;Wetzel et al., 2017)

| METHOD
The ContraVision approach presents the same topic from two opposite points of view (Mancini et al., 2010). To this end, we prepared two scenarios: a utopic video presenting the future of digital technologies in manufacturing in an optimistic light and a dystopic video having a negative take on the future use of these technologies. While we do not expect either of the scenarios to represent the future accurately, they provide a way of eliciting a broader range of responses from the participants than would have been expected through a single scenario (Mancini et al., 2010).

| Study design
The study adopted a mixed methods approach to address the research questions. The study used a between-subjects design to investigate the influence of DMT portrayal on public perceptions, without interference effects influencing responses to the utopic/dystopic scenarios. The portrayal was presented in two conditions: a utopic scenario and a dystopic scenario. Participants viewed either the utopic or dystopic scenario in the form of a video, each lasting about 3 min, followed by a questionnaire. Qualitative responses were examined via a thematic analysis (Braun & Clarke, 2006) to explore the reasoning behind attitudes toward the DMTs in question. The data were collected in November 2020 using an online survey distributed through the Prolific recruitment platform (https://prolific.co/).

| ContraVision video productions
Both fictional scenarios present a monologue by an imaginary manufacturing worker working for the local community factory in a small town. The full video recordings of the scenarios can be found in the University of Nottingham's research data repository (link: http:// doi.org/10.17639/nott.7176). Content and production of the videos were developed in collaboration with an independent film production company. The development process involved the team of researchers identifying a set of DMTs of interest within the Industry 4.0 paradigm, which included industrial and collaborative robotics, distributed data technologies/IoT, human augmentation and physiological sensing systems, and data visualization technologies. These DMTs were then considered from opposing perspectives, such that the two films were intentionally designed to mirror each other in terms of topic and content. We purposefully took near-future scenarios for these perspectives, employing a design fiction paradigm, such that they would be possible within current society, as opposed to, for example, extreme dystopian views portrayed in science fiction. The research team and film production company worked together in an iterative process to draft the script based on these considerations in such a way that the spoken narrative in both videos discussed the identified themes in the same order from opposing viewpoints, per the ContraVision technique's approach (Mancini et al., 2010). Three iterations of the script were made and reviewed by the research team before a voice actor was hired to provide the narration for both films. The audio was then overlaid over a selection of relevant visuals gathered from an industry reusable footage database to emphasize the points being made (although note: these visuals alone would not necessarily depict utopian or dystopian views without the narrative). The videos themselves also went through one iteration of feedback from the research team to make sure that the narrative communicated the key points clearly.

| Content of ContraVision narratives
Both scenarios begin by explaining that the factory has just undergone a period of significant change involving the integration of new digital technologies into routine work. The worker from the utopic scenario has a positive opinion about the changes, whereas in the dystopic scenario, the worker communicates his negativity and hesitation about speaking about this topic and making his opinions public. Overall, the main difference between the scenarios is the way in which the factory management has implemented the DMTs; in the utopic scenario (Figure 1), a user-centered, inclusive, and ethical design approach leads to direct benefits to workers, whereas in the dystopic scenario (Figure 2), DMTs have been introduced to oversee and control the work environment, leading to distrust, anger, and reduced productivity among workers.
Within the scenarios, the worker's monologue focuses on his perspectives on the introduction of several different DMTs, including wearable sensors for operator state monitoring, distributed data technologies, and industrial robotics. In the utopic scenario, the worker discusses having his heart rate and brain activity monitored for the sole purpose of improving wellness in the workplace while encouraging colleagues to help each other stick to wellness programs. In one excerpt, the protagonist states that the technologies are used to enable wellness, and that "we can even buddy up with a colleague to help each other stick to our individual wellness program. Paul Dixon and I are currently leading in our department." However, in the dystopic scenario, the same technologies are shown, but this time, they are not being used for the same purposes. The worker explains the same technologies are not being used out of an interest in the workers' wellness, but instead to track their levels of concentration, physical activity, and other metrics that are converted into a performance score. This is emphasized when the worker states, "We all get a score that takes into account all types of metrics. This score is also affected by how much we speak to each other about non task related things… We've noticed that no one is able to speak with [Paul Dixon] for more than a few seconds before the system moves him on." The advantages of distributed data and systems, coupled with virtual reality capabilities, are described in the next scene of the utopic scenario, where the worker proudly speaks of his hydroponic garden project that is being developed in collaboration with experts from around the world. Real-time translation capabilities as well as seamless integration of their models and data allow for distributed teams to easily collaborate. However, in the dystopic scenario, the same distributed data technologies are not put to such uses and instead are used by the factory in a way that lacks transparency. The worker expresses suspiciousness toward the promise that individuals will not be able to be identified from their data and that the data will not be shared, especially when workers learn that the company records their behavior outside of work.
In addition, the worker in both scenarios describes the integration of industrial and collaborative robots into their workplace. In the utopic scenario, the worker describes forming a connection with the robot he uses at work, even assigning it a name, saying, "When I'm working with my robot, I call him Bob, he is able to sense when I am losing concentration, but also when we are working well, he is able to F I G U R E 1 Still imagery from the fictional scenario depicted in the utopic video F I G U R E 2 Still imagery from the fictional scenario depicted in the dystopic video MARINESCU ET AL. | 187 anticipate my movements and we work as one together." Their interactions appear to be natural and beneficial for everyone involved. In the dystopic scenario, the situation is the opposite; the robot has not been adapted to interact effectively with the workers, and it turns out that the issues in human-system integration are negatively affecting the productivity score of our protagonist. This is described by the protagonist, saying, "We have a new process the robot and I are having trouble collaborating on. I think it learned with someone much smaller than me. Basically, it's trying to weld before I put a certain component in, which makes it really awkward to reach.
It ends up taking more time and it's killing my [productivity] score." The final discussion point in the scenarios focuses on sensor data, relating closely to the physiological/wearable sensors and distributed data technologies discussed previously. In the utopic scenario, sensor data are used by the company to assess the capabilities of workers, with the worker providing an example where a task was redesigned after data showed the original task created significant worker fatigue. In the dystopic scenario, the same data are used to compare workers in facilities around the world, where all compete against each others' productivity scores and no recognition is given for positive behaviors.

| Procedure
The online form introduced the research goals to the participants, and if they agreed to take part, they were redirected to a Microsoft Forms survey containing the informed consent form. Then, participants were asked to watch one of the video conditions and to complete a questionnaire consisting of both quantitative and qualitative questions. The questionnaire, shown in Table 1, was designed following a review of public perceptions questionnaires in other domains (Castell et al., 2014) and a discussion within the research team, which identified key areas of interest; from this process, questions were designed to collect information on individual attitudes toward the intersection of digital technology and manufacturing. These were piloted within the project team to ensure meaning was clear and addressed the concepts of relevance.
After the questionnaire was completed, participants optionally provided data on age, gender, level of education, and experience in manufacturing. Following this, they were redirected to the Prolific page for payment.

| Data processing and analysis
Out of the initial participants, 32 were removed based on the time they took to complete the survey. Participants who took less than 3 min plus the duration of the video to complete the questionnaire were determined not to have given the questions sufficient consideration, so these data points were removed; indeed, some participants actually took less than the duration of the video, indicating that they most likely did not watch most of it. After this removal, additional recruitment occurred to create equal-sized groups between conditions. In total, data from 134 participants were retained from a total of 166 submissions.
The data analysis consisted of an initial analysis of the quantitative data followed by an analysis of the qualitative responses. For the quantitative data, a Mann-Whitney U-test was used to determine if there were any differences for questions 1 to 14, between the utopic and dystopic conditions. In the case of the qualitative data, a thematic analysis was conducted in accordance with the approach set out by Braun and Clarke (2006). Guided by the thematic analysis, data were coded, the codes were examined, and where similarities were found sorted into themes and accompanying subthemes.

| Participant data
The analysis included data from 134 participants (61% females, 39% males), all older than 18 years (10%, 18-21; 26%, 21-29; 26%, 30-39; 18%, 40-49; 10%, 50-59, 8% 61+). An equal number of participants took part in both conditions, and the proportion in each age group was comparable between conditions. In terms of educational attainment level, the majority of participants had completed a degree in higher education (undergraduate or postgraduate level), while a subset of the sample held GCSEs, A-Levels, or equivalents. Figure 3 presents the breakdown of educational attainment within both conditions. Participants also represented a range of familiarity with manufacturing. When probed about their experience level in this field, where 1 represented "Never worked in manufacturing" and 5 "Highly experienced in manufacturing", a small proportion had some degree of experience (31.3%), whereas the majority had none (68.7%).
Participants were recruited using the Prolific platform and custom pre-screening was applied to select only participants from the United Kingdom. Participants were provided with £1.88 for an estimated 15 min of their time (the equivalent of £7.52/h); in most cases, it took less than 15 min to complete the study, and therefore the average pay was equivalent to £15.41/h. This study was approved by the University of Nottingham's Faculty of Engineering Ethics Committee. Question Scale The first 6 questions refer to the videos that were watched by the participant 1 How positive would you feel about working for this company? 0 (Very negative)-6 (Very positive) 2 Would you trust this company with your data? 0 (Not at all)-6 (Fully trust) 3 If you worked in a manufacturing company, do you think such technologies would make your job easier?
0 (Not at all)-6 (Very likely) 4 If you worked for such a manufacturing company, would you be happy to wear sensors collecting physiological data?
0 (Not at all)-6 (Very happy) 5 If you worked for such a manufacturing company, would you be happy to work alongside a robot?
0 (Not at all)-6 (Very happy) 6 If you worked for such a manufacturing company, would you see the benefits of distributed data technologies?
0 (Not at all)-6 (Yes) The following questions refer to digital manufacturing technologies in general 7 Overall, how positive or negative do you feel about digital manufacturing technologies? 0 (Very negative)-6 (Very positive) 8 Overall, how positive or negative is the impact that manufacturing has on: 1. The UK as a whole 0 (Very negative)-6 (Very positive) 2. Your local community 0 (Very negative)-6 (Very positive) 3. Your family 0 (Very negative)-6 (Very positive) 4. Your individually 0 (Very negative)-6 (Very positive) 9 Overall, how positive or negative is the impact that digital technology has on: 1. The UK as a whole 0 (Very negative)-6 (Very positive) 2. Your local community 0 (Very negative)-6 (Very positive) 3. Your family 0 (Very negative)-6 (Very positive) 4. Your individually 0 (Very negative)-6 (Very positive) 10 Digital technology has an important role to play in meeting the challenges the UK faces 1 (Strongly disagree), 2 (Tend to agree), 3 (Neutral), 4 (Tend to agree), and 5 (Strongly agree)

| Personal data
The theme of "personal data" emerged during the analysis, representing attitudes that participants held toward the individual employee's personal data, which is to be collected by the systems in the utopic vision. Subthemes were identified as relating to attitudes toward data "capture" and "data usage." Participant feedback indicated an interest in the topic of data capture. One quote exemplified concern about this topic, particularly regarding the employee monitoring in the workplace, stating, "What kind of dystopian ideal is this? Forcing workers to be biometrically monitored during their time at work?" (Participant 16U). This quote relates to the scenario presented in the video in which data on employees were captured throughout the day, whether they were at work or not. As presented in the videos, monitoring in real-world environments ranged from evaluating productivity to physiological health-related assessment for individuals, the latter referred to in the quote. This subtheme reveals concerns with the notion that capturing physiological data about employees is "a dystopian ideal" or that certain forms of it are less palatable than others; for example, one participant stated that they were, "OK with heart and breathing rate being recorded, [but I] find brain activity a bit creepy" (Participant 54U).
Participant responses also indicated concerns related to how personal data is used and shared, captured in the "data usage" sub- Second, there is an assumption that this personal data will be analyzed on an individual basis. It may be possible that respondents' attitudes would differ if data were aggregated over a large data set.

| Impacts to people
The next theme identified was "impacts to people," concerning individuals as a collective rather than specific individuals or the wider society. This can be evidenced by the subthemes of possible discrimination and the perceived degradation of human contributions in manufacturing workplaces.
Within this theme, the first subtheme focused on concerns related to "potential discrimination" against workers. Participants picked up on the idea that the certain workers may not be able to acquire work if their performance is being assessed. There are overlaps here with the concerns expressed in the "personal data"' theme, in that the concept of data runs throughout, but differs in that the discrimination aspect looks at the broader implications of data use. showing little emotion. This is echoed in another participant's quote, which stated that, "It scares me that we are becoming a society that will depend mainly on data. I fear that it will remove any connection to reality" (Participant 49U). The data referred to could also be linked to the degradation of human values, in the sense that rather than making personal decisions, we start making all decisions based on data.

| Future
The final theme that emerged from participants in the utopic condition Two additional subthemes emerged relating to a "positive outlook" versus a "negative outlook" on future technology. These subthemes may seem mutually exclusive, but there is an argument that some participants held mixed feelings toward the technologies. With regard to the "positive outlook" subtheme, one participant commented on the potential for DMTs to affect individuals by stating it was "amazing how it can improve productivity" (Participant 42U). While it is difficult to ascertain from the participant's comment whether the technology's perceived benefits to productivity are within or outside of work contexts, other responses suggest a general positive attitude toward the potential benefits that digital technologies can bring to individuals.
In contrast, the "negative outlook" subtheme related to how technology is perceived in a negative way as it moves into the future.
One participant questioned the consequences of moving toward technological advancement shown in the video, saying, "[…] i wonder how much robots impact people's jobs, especially when so many more people have been made redundant" (Participant 31U). One idea from this to consider further is the use of the term "wonder," which presents a slight uncertainty of the future, with the comment around job losses adding a negative context to this uncertainty. Further, a negative comment can be found within the next quote: "I think this technology will be used to exploit workers rather than improve their workplace. Why would technology make the workplace better if the people in charge have the same ethics codes and interests (i.e., profit) as before?" (Participant 47U). It is difficult to be sure what element of the technology will be used to what end, but there is this perceived sense that the technology will generally be used in a negative way.

| Impacts on wider society
One theme that emerged from participant responses in the dystopic condition was attitudes and concerns around the "impact on wider society." Participant responses in this theme related to attitudes toward the broader lens of society, rather than on perceived impacts at the individual level.
The first subtheme revolved around a "perceived positive impact for industry." Compared to participants in the utopic condition, dystopic condition participants assumed a perspective that was more industry- concerned that "… it eliminates human interaction and places robots in place of the workers" (Participant 65D). Comparing the two quotes, a common strand is a concern that increased technology will reshape the social dynamics of future workplaces, putting greater value on contributions of technology, and reducing sources of wellness for human workers like interpersonal interaction.
Lastly, concerns surrounding "job security risks" emerged from dystopic participant responses, with comments relating to the participant's attitudes toward the impact of DMTs on job security.
One participant voiced concern over ethical implementation and employee safeguarding, stating, "[…]but it is also important to introduce them in the right way to avoid mass redundancies without other jobs being available or a universal basic income like alternative" (Participant 21D). Concern for job security is further evidenced in another quote where the participant felt it was important to "[learn] how robots and humans can work together so that humans don't [lose] their jobs to robots" (Participant 60D). This comment further reinforces the idea job security is a potential issue for DMT.

| Attitudes toward technology versus ContraVision portrayal
In our second research question, we sought to identify the effects of how the DMTs portrayed in the videos influenced attitudes toward them. To this end, a series of Mann-Whitney U-tests analyzed each of the Likert scale questionnaire items to determine if there were differences in ratings between the utopic and dystopic conditions; results are shown in Table 2. In addition, for each one of the questions, the distributions of ratings for both conditions were assessed by visual inspection. As shown in Table 2 (Bitkina et al., 2020).
Design and organizational factors influence attitudes toward future systems; for example, factors including cost, usability, trustworthiness, flexibility, cost, and learnability have been identified as critical challenges in the design of virtual human factors tools such as digital human models and virtual reality (Perez & Neumann, 2015).
Within the field of HFE, frameworks such as the Technology Acceptance Model (TAM; Davis et al., 1989) and the Automation Acceptance Model (AAM; Ghazizadeh et al., 2012) are wellestablished in the literature for predicting the successful adoption of a new product or system. The original TAM predicts acceptance via actual system use, which is moderated by an individual's intention to use the technology, which in turn is influenced by attitudes toward the system, perceived usefulness, and perceived ease of use.
In an extension of the original model, the AAM incorporates the moderating effects of compatibility, trust, and external variables on the TAM constructs (Ghazizadeh et al., 2012). As we have demonstrated, ContraVision can be used to delve into factors influencing technology acceptance before actual system use is possible. Although we did not structure the current study to explore acceptance specifically, our findings provide insight into TAM-related constructs such as perceived usefulness and attitudes toward using Previous work on the acceptance of DMTs, such as with distributed data technology, has found that while such systems can benefit organizations and individual users, technical and social challenges must be addressed, particularly with regard to ensuring trustworthiness, privacy, and data security (Atzori et al., 2010;Fast & Horvitz, 2017;Zubiaga et al., 2018). In the current work, participants voiced fears that personal data could be used by employers to discriminate against workers, creating work environments more focused on evaluating workers based on sensing and remote observation (e.g., via the physiological sensing-based operator state monitoring system shown in the videos) rather than on demonstrable behaviors and performance. Furthermore, some technologies were seen as potentially invasive, posing risks to individual privacy and security. In a survey of European service organizations on IoT security, Asplund and Nadjm-Tehrani (2016) observed that service availability was prioritized more highly than data confidentiality. As DMTs continue to mature and be integrated into workplaces, organizations and regulatory bodies must address the potential pitfalls and misuse of personal data; ongoing work in this area includes the development of "right-to-know legislation" requiring systems to show the user what data is collected about them and providing the option for an individual to remove their data (Weber, 2010).
Although the present research shows that members of the UK public hold concerns about DMT data capture and its impact on individual lives, longitudinal research indicates that perceptions toward technology evolve as systems mature. Zubiaga et al. (2018) conducted a longitudinal analysis of social media posts related to IoT technology, finding that public perceptions toward trust, security, and privacy grew more positive over the analyzed timeframe. The authors identified specific concerns related to IoT technology and related concepts, including analytics, machine learning, big data, security implications, and machine-to-machine communication. Over the time period, posts increased in positivity toward analytics and machine learning topics but became significantly more negative toward security topics. Similarly, in a longitudinal analysis of public reports on AI, Fast and Horvitz (2017) found that ethical concerns related to AI were appearing with increasing frequency in public dialogs. These findings align with the present study's findings, where some participants reported generally positive views toward DMTs and digital technology in general but also expressed concern for ethical, secure, and fair use of data, particular types that are personal in nature.

| Maximizing business benefits while minimizing negative impacts on job security and wellness
Throughout history, new technology has frequently led to cultural anxiety surrounding its introduction into the workplace; concerns have namely related to machines replacing human workers and a resulting degraded quality of life for workers (Mokyr et al., 2015), concerns that were reflected in the current study. Records dating from the 18th century through to the modern era show that economists and policymakers have long debated the potential implications of new workplace technology on society, but that many predicted impacts have not been fully realized (Autor, 2015;Mokyr et al., 2015). While the emergence of advanced digital technologies such as AI and sensing have generated familiar predictions, scholars suggest that the functions performed by such systems may change the nature of work, but that humans will still play a significant and meaningful role in such work systems (Mokyr et al., 2015). Similarly, challenges associated with job security and individual wellness became apparent from participant responses. Distinct from concerns about ethics, privacy, and security, responses indicated that participants held concerns that DMTs could reduce or remove the need for human workers. In the case of operator state monitoring technologies, participants were also concerned that these could be used to discriminate or micromanage workers if not managed responsibly. This is of significant interest given the recent rise in interest in sensing technology for improving operational safety by tracking parameters such as mental workload (Argyle et al., 2021;Marinescu et al., 2018), situation awareness Zhang et al., 2020), and fatigue (Sikander and Anwar, 2018); based on participant responses, we argue that research should focus not only on developing functional technologies but that design and implementation should assume a user-centered design paradigm, focusing on creating systems that result in safer operations and an engaged, satisfied workforce. These concerns link closely with a challenge for the industry to ensure that DMTs have positive impacts on worker wellness, holistically considering both physical health and mental well-being. In line with this, previous work has observed increasing levels of fear surrounding the loss of control over technologies used to inform critical decisions and decreasing levels of positivity toward the impact of AI on human work (Fast & Horvitz, 2017). In relation to distributed data technologies, organizations have responsibilities toward their employees to maintain data security; with regard to the gathering and analysis of personal data, allowing users to have a degree of control over these processes can enhance trust in such systems (Maple, 2017).
In light of these concerns, we suggest that DMT integration programs consider not only how to facilitate the active participation of users but also how to communicate the potential values of the technology in personalized ways to fit the target audience. For example, previous research has shown that workers with different backgrounds will consider different factors when whether or how to use workplace technology (Morris et al., 2005). In an investigation based on the Theory of Planned Behavior (Ajzen, 1985), Morris et al. (2005) found that workplace technology adoption perceptions varied, identifying significant gender differences as participant age increased, but not among younger participants. Furthermore, among the participants in the older age group, the authors found that men's perceptions toward adopting a particular technology were most influenced by their attitudes toward it, incorporating perceived usefulness and positive/negative perception of using the tool. In contrast, women in the older age group held attitudes that were most influenced by perceptions surrounding social norms, ease of using the technology, as well as an attitude toward the technology. Although we were not able to explore gender or age differences in the current study, these findings provide insight into addressing societal challenges around ensuring that DMTs are designed and implemented in a user-centered manner, accounting for differences in worker characteristics that may affect intentions to use the systems.
It is important to note that, despite long-held cultural anxiety around new technology, digital technologies have the potential to enhance work flexibility and quality of life, two aspects that were limited during previous industrial revolutions (Mokyr et al., 2015).
Recent research on the Millennial workforce has suggested that individuals value work that offers flexible work patterns, challenging tasks, roles of responsibility, and opportunities for professional development (Schaar et al., 2019). We hypothesize that communi-  (Castell et al., 2014), and further research is needed to explore the impact of educational and public awareness campaigns on perceptions of DMTs in subject matter experts and the wider public. Public perceptions of new technology tend to be an understudied area (Bellamy, 2019), but exploring the factors that shape perceptions may provide a valuable starting point for recruiting and training future digital manufacturing experts. This is especially important as technology acceptance research tends to focus on technology adopters, often with limited input from those that have not yet adopted a technology (Verdegem & De Marez, 2011). In a synthesis of five major technology perceptions models, Bellamy (2019) identified that beliefs about technology were a function of four multifaceted dimension: "knowledge of technology," "project scope," "impacts of technology," and "trust in the control of technology." Within the proposed toward the target technology, shaped by the viewpoint to which individuals were exposed (Mancini et al., 2010). Furthermore, the technique provokes thought on topics that may be difficult for participants to identify with due to a lack of prior experience. Here, this was due to the futuristic nature of the given scenarios and DMTs.
Analyses comparing responses between participants in the utopic condition and the dystopic condition supported previous findings in this regard.
Given that the videos presented such different visions of the future, it was expected that a significant difference in opinion would be found between the two conditions. However, two aspects of the responses were particularly interesting to note: the distribution of the attitudes between Utopian and Dystopian conditions, and the questions for which there were no significant differences, despite descriptions in qualitative feedback. When probed on their attitudes toward the company and DMTs mentioned specifically in the videos, participant responses between the two conditions were significantly different. When asked how positive they would feel about working for such a company, participants in the dystopic condition responded highly negatively, as expected; however, while a symmetry in ratings for the utopic condition might have been expected, this was not the case, and the median ratings reflected only a slightly positive attitude.
This echoes the thematic analysis, which showed that even when presented from a utopic perspective, participants still held concerns over the misuse of such technologies.
A similar phenomenon was observed when participants were asked whether they would trust this company with their data. Again, as expected, those in the dystopic condition responded highly negatively while there was more of a uniform distribution across ratings in the utopic condition, with a median of 3, indicating a neutral attitude. These findings were also in line with the thematic analysis results for the utopic condition. Responses also revealed that when presented with the Dystopian view, participants slightly disagreed that the technologies in question would make their manufacturing job easier, but those exposed to the Utopian view believed the impact on their work would be largely  (Sandhu et al., 2020), workshops (Perez & Neumann, 2015), and semistructured interviews (Asplund & Nadjm-Tehrani, 2016). In addition, methods such as Cognitive Work Analysis (Salmon et al., 2016), Event Analysis of Systemic Teamwork (Walker et al., 2006), and Ecological Interface Design (EID) (Vicente & Rasmussen, 1992) have been widely used to analyze properties, interactions, and emergent behavior within complex sociotechnical systems to inform the design of future technologies. Kant and Sudakaran (2022) proposed an extended approach to EID, the integrated EID (iEID), motivated by the need for novel design approaches for increasingly digital systems and validated this in the design of a digital twin. Within iEID, early-stage activities involve developing a conceptual model of the context of use, with the recommendation to capture data on the work domain, tasks, situations, and operator characteristics through qualitative analysis.
It is in HFE endeavors such as this that we believe the ContraVision technique would have the most significant impact. ContraVision is a useful tool for exploring aspects of the design before a technology or system is fully realized, and as we have demonstrated through the questionnaire-based study, the design fiction can provide relatable narratives that provoke thoughts on a range of factors. As shown through the thematic analysis in particular, participants provided insight into attitudes toward the context of use and interactions within the fictional future manufacturing system, two aspects considered essential in the systems ergonomics perspective (Wilson, 2014). As the ContraVision films provided a holistic view of the DMTs, it is possible that this holism provided enough understanding for nonexpert participants to identify with, allowing them to think through complex design questions in meaningful ways.

| Limitations and future work
There were several aspects that limit the conclusions that can be drawn from this work, primarily related to sample size, sample demographics, and the nature of the data collection instrument. First, the study focused exclusively on attitudes held by residents of the United Kingdom, so the results may not generalize to other countries and cultures; indeed, previous research has shown that there are cultural and societal differences in perceptions of technology (Muk & Chung, 2015;Zubiaga et al., 2018). Second, the sample was skewed toward younger, highly educated participants, with little prior Participant feedback clearly indicated the need for secure, trustworthy systems that protect personal data rather than exploit it to the detriment of workers. Additionally, responses revealed the importance of clearly communicating the value of and potential benefits of DMTs while recognizing that different end-user groups may have different needs that must be understood and accommodated during the technology design and implementation process. Furthermore, participants in both utopic and dystopic conditions voiced concerns surrounding being forced to use technology to benefit the fictional organization but not the individual. This points toward risks with the balance between end-user engagement and satisfaction, and based on these results, we argue that implementation programs should consider whether tools and technologies serve the interests of both the employer and the employees, moving from a "technology for technology's sake" approach and toward a paradigm that considers and incorporates a diverse range of stakeholder values and feedback.
Second, based on this demonstration, we argue that the ContraVision technique offers a valuable framework for eliciting user feedback on technologies and systems that are difficult to interact with directly, something that complements existing HFE methods and is of value in systems ergonomics research. Although the future of manufacturing is unlikely to be either fully utopic or fully dystopic, by understanding and addressing societal concerns about these systems, we can provide direction to enable not only more productive and efficient workplaces, but also workplaces that are more equitable and acceptable to their workforce.