Introduction

An increasing number of cities are developing towards, or harboring ambitions to become, Smart Cities. While the definitions of a Smart City vary, many of them include the use of technology and sensors for ubiquitous data collection (Albino et al., 2015; Goodman, 2020; Kitchin, 2016; Kondepudi et al., 2014). The aims of this increasing digitalization and data collection are manifold: They range from increasing efficiency of public services to serve a growing population, to better prediction and handling of natural disasters, to increasing energy efficiency, or enhancing quality of life (Kondepudi et al., 2014).

Yet, as these efforts and the related interventions require linking individual people (their actions, locations, preferences) with information about the functioning of the city more broadly, these interventions are not without ethical implications. While often intended for good, ethical dilemmas can result from the tension between collecting potentially sensitive data from individuals to achieve benefits on a societal level. Examples include information-providing interventions such as smartphone warning apps (Reuter et al., 2017), e.g. COVID-19 warning apps that send warnings based on information provided by other users. Other interventions focus on encouraging certain user behaviors, such as interfaces supposed to encourage individual energy saving (D’Oca et al., 2014). And finally, other interventions make use of personal data to trigger certain actions. Examples include the use of sensor data to manage traffic flows (Po et al., 2019).

As such, privacy concerns resulting from the ubiquitous collection of user data are among the most pressing ones mentioned by Smart City inhabitants (Emami-Naeini et al., 2023) and debated in the current Smart City literature (Juvenile Ehwi et al., 2022; Silva et al., 2018). For example, Ziosi et al. (2023) illustrate privacy concerns stemming from data collection as follows: “the technologies involved in a smart city do not just make the trains run more efficiently, they also enable city officials to collect information about the train schedule and train passengers with techniques including facial recognition scans, gait recognition, body temperature, and more” (p.12).

People express a particular concern when data have been collected without their explicit awareness. This can even lead to the abortion of projects (Nam & Pardo, 2011), such as the case of video footage of smart streetlights used for surveillance without disclosing this to pedestrians (Figueroa, 2020; Stegall, 2020). Another example for ethical concerns in Smart Cities based on interventions being conducted unawares is provided by nudges. These are small interface tweaks supposed to encourage a “wise” choice (Thaler & Sunstein, 2008). Based on their mode of action, which makes use of automatic cognitive processes such as biases and heuristics, and depending on their level of transparency, nudges have the potential to be manipulative thereby limiting users’ freedom of choice and autonomy (Hansen & Jespersen, 2013). An example for data-driven nudges is provided by the city of Eindhoven that tries to reduce nightlife crime by influencing the citizens behavior through altering the street light color and intensity. By adjusting the streetlight to predicted mood the aim is to make people calm down or take different routes (Hoogeveen et al., 2018). While well-intended, interventions like these could be perceived as manipulative if pedestrians are not aware of the intervention or its intention (Ranchordás, 2020).

Reflecting on the ethical implications of interventions conducted in Smart Cities is especially relevant given that the Smart City context can serve as a kind of testbed for rolling out and testing interventions involving humans. These interventions can be led by researchers studying human perception or behavior, by industry testing new technologies or services, or by civil servants evaluating the effectiveness of certain urban policies. In the past, this was often limited to one-off or occasional measurements, limited in scope and locally bound, e.g., to a university laboratory or certain city area in a field study (Kitchin, 2016). Examples mentioned by Kitchin (2016) include censuses, household surveys or commissioned interviews that provide a snapshot of particular aspects of city life at a certain point in time. In contrast, reaching or gathering data from a multitude of people is now easier than ever given the interconnectedness of devices and sensors within the Smart City context. The generated big data, e.g., data from all train travel, is found to have fundamentally different characteristics as compared to previous small data sets, e.g., survey data from a few train travelers (Kitchin, 2016). This raises new challenges for implementing, testing, and evaluating Smart City interventions that involve humans or human-generated data in an ethical way.

For research and practice involving humans, psychological associations worldwide have established ethical guidelines that commonly incorporate protections enshrined within the Declaration of Helsinki (American Psychological Association, 2017; European Federation of Psychologists’ Association, 2005; The British Psychological Society, 2012). Psychology as a discipline studies human perception and behavior. As such, psychology cannot be applied in and by itself, but needs to be considered in context, such as an educational, clinical, social, or organizational context. To account for that, psychological ethical guidelines have been developed so that they are applicable to various contexts in which humans decide, act, or learn. Therefore, in theory they should also be applicable to humans living and acting in a Smart City context.

Many other disciplines and associations provide ethical guidelines as well. Some concern certain technology, such as robots or artificial intelligence (AI). However, Smart Cities include a multitude of technologies and evolve with technological developments so that the application of guidelines designed for a specific technology can perhaps not account for the dynamic and complex nature of Smart Cities. Other ethical guidelines concern concrete application areas, such as medical guidelines or guidelines applied in a military context. Again, while these are highly valuable within their application areas, they might be too specific to account for diverse interventions and scenarios within Smart Cities. Therefore, instead of a certain technology or application area, this article focuses on the humans within Smart Cities. And for interventions that concern humans, i.e., their perception, behavior, attitude, motivation, or decision-making, psychological ethical guidelines provide a well-founded starting point. In addition, McMillan et al. (2013) and Renaud and Zimmermann (2018) have already shown that these guidelines are flexible enough to be applied to large scale trials of mobile software as well as digital privacy and security nudges, respectively.

In a similar manner this article will therefore trial and apply existing psychological guidelines for ethical research and practice involving humans to the Smart City context. The aim is to reflect on their applicability to Smart City interventions, to identify ethical challenges, and to reflect on potential gaps as well as avenues for further research. The aim is to provide a helpful concretization of ethical aspects relevant especially for the Smart City context and to spark further research on human-related Smart City interventions.

Related work on ethics in smart city contexts

Starting in as early as the 1990s and with a significant increase in 2018 and 2019, there is a huge body of literature on Smart Cities (Zhao et al., 2021). Yet, despite multiple articles and literature reviews concerned with the definition of the Smart City (Chourabi et al., 2012; Mora et al., 2017; Yigitcanlar et al., 2018; Ziosi et al., 2023) a distinct conceptualization is still missing. For example, Yigitcanlar et al. (2018) instead uncovered a multi-dimensional Smart City framework with three drivers, i.e. community, technology, or policy, and desired outcomes, such as productivity, sustainability, or wellbeing. Therefore, the next section takes a closer look at existing definitions and the special characteristics of Smart Cities to establish a common ground for the application of ethical psychological principles.

Smart Cities are an interdisciplinary field of research and have been analyzed from various perspectives. For example, Kummitha and Crutzen (2017) provide an evolutionary perspective and propose the 3RC framework that differentiates Restrictive, Reflective, Rationalistic and Critical schools of thought. The restrictive school appears technology-driven with little mentioning of social aspects such as social inclusion or justice. At its extreme this technological optimism is even labelled techno-solutionism (Morozov, 2013). The reflective school picked up criticism related to the restrictive school and incorporated human elements such as enhancing human capacities through technology while still being technology-centered. The rationalistic school of thought sees a shift towards community-driven Smart Cities and a focus on the people in a more systemic approach. Critical schools of thought take a more pessimistic stance towards the Smart City concept and its practice. Concerns, e.g., include the privatization of urban space, data security, or Smart Cities as a business model. In line with that Hollands (2008) argues that beneath the more “humanist” emphasis on human capital and social learning in Smart Cities often lies “a more limited political agenda of ‘high-tech urban entrepreneurialism’“ (p.314). He suggests a progressive Smart City that seriously starts with the people and their interactions instead of technology-driven approaches. Following Hollands (2008) and the 3RC framework, this article takes a human-driven approach towards Smart Cities by discussing psychological ethical principles that put the human in the center. However, the discussion of the principles’ applicability includes the range of existing schools of thought by reflecting on the implications of a technology-driven approach for ethics, including concerns and criticism raised in the literature, and discussing the potential of community-driven Smart Cities to mediate these concerns.

In terms of research content, Zhao et al. (2021) or Chourabi et al. (2012) uncover prominent research areas and critical factors, such as planning and governance, technology diffusion, or people and communities. The ethics of Smart Cities often seem to only be implicitly considered in identified challenges such as accessibility or privacy of personal data (Chourabi et al., 2012). Hence, Zhao et al. (2021) call for more research on the social integration of technology, e.g. in terms of equality, accessibility and inclusion. Yet, several researchers explicitly focus on the ethics of Smart Cities. For example, Juvenile Ehwi et al. (2022) suggest that while much of the ethical Smart City debate focuses on data handling, it should be extended to the ethical underpinnings of decision-making as part of Smart City governance. Chang (2021) identifies four emerging issues for an ethical Smart city framework that include privacy/surveillance and data integrity but also social equity and the aging population and public transportation. Ziosi et al. (2023) see the need for ethical scrutiny based on Smart Cities not only addressing traditional city problems and providing new opportunities, but also new challenges. They describe four dimensions of ethical concerns in Smart Cities that range from privacy concerns related to the network infrastructure over post-political governance and social inclusion to sustainability aspects. An additional challenge is that not all impacts are easily visible, such as potential shifts of responsibility for neighborhood safety from city officials to traffic control rooms with increased traffic surveillance (Ziosi et al., 2023). Kitchin (2016) also uncovers several ethical challenges especially related to big data privacy. He calls for responsible research beyond laws and “the adoption of ethical principles designed to realize benefits of smart cities and urban science while reducing pernicious effects” (p.1). Kitchin (2016) suggests that professional bodies review their ethical standards in light of big data. Based on that, this review of the ethical psychological principles in the context of Smart Cities can be seen as a first step towards exploring their current applicability, uncovering ethical challenges resulting thereof, and exploring potential for revision and supplementation with additional principles.

Characteristics of smart cities

As introduced above, the term Smart City concerns a complex concept with no agreed upon single definition (Camero & Alba, 2019; Yigitcanlar et al., 2018). However, work that falls under the umbrella of Smart Cities bears several similarities:

First, one main idea behind the Smart City concept is to increase the efficiency of urban operations and services, and their competitiveness (Etezadzadeh, 2020; Silva et al., 2018; U4SSC, 2021).

Second, many definitions stress sustainability aspects alongside efficiency (Kondepudi et al., 2014; Srivastava & Sharifi, 2022), i.e., ensuring that “the needs of present and future generations are met in the full respect of economic, social, environmental as well as cultural aspects and values of society” (U4SSC, 2021). The standard ISO (2019) “Sustainable cities and communities — Indicators for smart cities” even focuses on the sustainability aspect in the smart city definition by stating that a smart city “increases the pace at which it provides social, economic and environmental sustainability outcomes”.

Third, various definitions agree on the use of information, communication, and sensor technologies to achieve that aim. Thereby, the increasing digitalization, the interconnection of systems and information as well as the use of ubiquitous sensor data play a major role (Albino et al., 2015; Goodman, 2020; Kitchin, 2016; Kondepudi et al., 2014).

And finally, a number of definitions explicitly mention the involvement of or focus on the human, e.g., “in brief, technologies and tools need to be put at the service of the people (p.3)” (U4SSC, 2021). Other definitions furthermore include the aspect to enhance quality of life and benefit the inhabitants (European Commission, 2023; Kondepudi et al., 2014; Silva et al., 2018; U4SSC, 2021). In line with that, the definition provided by the OECD (2019) also includes the aim for cities to enhance active engagement with their citizens.

The application of the Smart City concept focuses on critical infrastructures with a high relevance for the public good, such as traffic and transport, energy supply or health (Federal Office of Civil Protection and Disaster Assistance, 2022). Based on the aspects reviewed above, this article considers the following Smart City working definition:

Smart Cities aim to increase operability and efficiency of city operations and services, while maintaining and increasing sustainability of economic, social, and environmental aspects, by making use of information, communication, and sensor technologies, in a way that respects, involves, and directly serves the human.

This definition entails several characteristics of Smart Cities that impact the reflection on ethical guidelines for Smart City interventions and pose challenges for their application. The first is that the Smart City concept is located at the interface of technological and social dimensions as illustrated in Fig. 1 (OECD, 2019). As such, Smart Cities are socio-technical systems. Meijer and Bolívar (2016) clustered smart city publications into those with a focus on technology use, on humans or on their interplay in a socio-technical perspective. While they found previous research to mainly focus on technical aspects, the socio-technical perspective is found to be the richest perspective for analyzing the socio-technical dynamics of a smart city.

Fig. 1
figure 1

Smart Cities at the intersection of technology and people

Full size image

Furthermore, across definitions, a common concept is the ubiquitous collection and use of data, e.g., through sensors. This involves the collection of sensitive data, e.g., personal data such as personal preferences, location data, or biometric data. Next, the increased interlinkage of people, devices, services, and data leads to high degree of complexity and interconnectedness (Kitchin, 2016). This again causes that, first, system outcomes cannot be easily traced back in a simple, linear chain of cause and effect to a root cause. Instead, system outcomes are influenced by various factors that interact in a non-linear fashion so that system outcomes are emergent. Second, even small events, triggers or interventions can cause unintended cascading effects. Third, the interconnectedness makes it difficult to determine the boundaries of the socio-technical system, to isolate effects, and to limit interventions, e.g., to a single individual and a device, within a certain unit of an organization, or within the target group only. System boundaries become increasingly blurred leading to an unclear reach of interventions. Consider the following example shown in Fig. 2 in four stages (a) –(d) to illustrate the Smart City characteristics and the challenges resulting thereof:

Fig. 2
figure 2

Graphical illustration of a human-centered intervention applied in the Smart City context that produces cascading effects and emergent, unexpected outcomes and even outside the assumed system boundaries of a certain city

Full size image
  1. a.

    Imagine the city administration of a fictional Smart City teamed up with local researchers to develop an intervention that encourages people to save energy. As such, the intervention targets human cognitive processes and behaviors. The intervention’s aim is to increase the city’s energy efficiency and strive towards global sustainability goals.

  2. b.

    The city administration and the researchers develop an intervention that uses nudges to target social norms, e.g., by displaying an average level of energy consumption as compared to one’s actual consumption, and social comparisons with neighbors and friends. The intervention builds on previous research that shows how social influence motivates certain behaviors (Caraban et al., 2019; Torning & Oinas-Kukkonen, 2009). The intervention is primarily targeted at the adult population of the city and contains information for house or flat owners, e.g., on heating and insulation options. To reach many citizens, the intervention is deployed as a smartphone application that is advertised via the city’s website and available via the app store. To measure energy consumption, the app asks users to provide information on energy use and housing situations and allows for integration with the data of local energy providers.

  3. c.

    The first citizens provide energy data, interact, and compare results. Initially, the intervention produces the desired effect and motivates the citizens to reduce their consumption visible from the data provided in the application and the local energy provider. Through personal and digital communication, the smartphone app spreads further and reaches people outside the targeted city, such as acquaintances, relatives, or colleagues of the originally targeted citizens.

  4. d.

    The new users of the app may be different from the group initially targeted. They might be from a larger city with a higher percentage of tenants as compared to homeowners. For them, information on heating and insulation options might not be relevant or even frustrating. At the same time, they might live in smaller, but newly built flats with modern energy standards. Their average energy consumption levels might thus already be lower as compared to an average user. These complex interrelations could lead to a paradoxical increase in energy consumption based on a tendency towards the average not anticipated by the app developers and similar to the effect described by Schultz et al. (2007). Additionally, unexpected challenges may arise from the increase in users and diversity, e.g., support for elderly people or a translation to different languages might be required.

This example illustrates how the Smart City characteristics impact on the ethical implementation of interventions involving humans. As an example, the intervention has been decided upon by the city administration and researchers but did not include citizens as affected stakeholders. The smartphone application collects potentially sensitive information on energy consumption and shares it with other app users and local energy providers for comparison. The application uses social norms nudges, perhaps without the users being aware of its influence. Furthermore, app usage cannot be easily limited to the target user group in one city but spreads beyond assumed system boundaries and causes unanticipated and even counterproductive effects. The next sections thus first briefly introduce and then review psychological ethical principles. For each principle, the example of the energy-saving intervention described above will be reflected upon to illustrate how one intervention touches multiple if not all the principles.

Guidelines for ethical psychological research and practice

Previous work by McMillan et al. (2013) and Renaud and Zimmermann (2018) compared and summarized the ethical principles as proposed by different established psychological associations such as the American Psychological Association (APA) (2017), the European Federation of Psychologists Association (EFPA) (2005) or the British Psychological Association (BPS) (2012). The following builds on the five principles as described in further detail in Renaud and Zimmermann (2018):

Respect

The BPS (2012) describes the principle as follows: “Respect for the dignity of persons and peoples is one of the most fundamental and universal ethical principles across geographical and cultural boundaries, and across professional disciplines. It provides the philosophical foundation for many of the other ethical principles.” As the principle hence forms a basis for other principles, it is mentioned here first. The respect principle concerns several aspects. One is to respect individual differences and to not engage in discriminatory practices, e.g., based on gender, age, or race. A second aspect concerns the respect of the person’s autonomy and privacy, e.g., through careful anonymization.

Beneficence

As envisioned by the APA (2017) the beneficence principle is aimed at ensuring that research and practice is aimed at contributing to the good of the affected people. Furthermore, emphasis is placed on ensuring that people are protected from potential harm or risks, i.e., ensure nonmaleficence. Likewise, potential conflicts are to be dealt with in a way that harm is minimized. A final aspect mentioned by the APA (2017) is that safeguards should be taken against factors that can lead to misuse of influences and undertaken actions.

Justice

Justice as an ethical principle means to ensure equal access to and benefit from the intervention or measure for all persons (American Psychological Association, 2017). This includes considerations of fairness and ensuring equal quality of interventions and processes for all people. The principle furthermore suggests not imposing undue burdens on certain groups of people.

Scientific Integrity

For all interventions and measures deployed scientific quality and integrity should be ensured. For example, established scholarly standards are always adhered to. The EFPA (2005) and BPS (2012) add that psychologists should only offer services and apply methods within their area of competence, and also be aware of the limits of their competencies. While the principle has been described for psychologists, the content can be transferred to other disciplines and implicitly suggests interdisciplinary approaches involving people with different sets of competencies to address questions exceeding the limits of one discipline as often the case in Smart City environments.

Social responsibility

The principle social responsibility suggests that professionals should accept “appropriate responsibility for what is within their power, control or management” and to be accountable for one’s actions (The British Psychological Society, 2012). The APA (2017) even adds the aspect of caring for other professional’s compliance with ethical standards. It is furthermore aimed at fostering awareness for and reflection on the consequences, long-term effects, and potential side-effects of the deployed measures or interventions. The principle explicitly not only concerns affected individuals but society on a broader level.

While the principles have been developed with a focus on psychological research and practice, they are relevant beyond psychological trials or psychologists as a target group. Psychology as a discipline is concerned with human perception and behavior, and it is important (if not already obvious) to note that Smart Cities are built for perceiving and behaving humans. As humans are at the center of Smart City interventions, whereas certain application areas or technologies differ and change over time, the principles for psychological research and practice will be reviewed and discussed as starting point for informing the design, implementation, and evaluation of Smart City interventions of various kinds.

Application of ethical principles to Smart City interventions

The following subsections consider the applicability of the ethical principles to Smart City interventions and elaborate on challenges that stakeholders may encounter when designing, implementing, or evaluating Smart City interventions that involve humans. Figure 3 summarizes the challenges discussed in the following subsections related to each principle with slight overlaps of challenges that are related to two different principles.

Fig. 3
figure 3

Summary of Exemplary Discussed Smart City Challenges related to the Ethical Psychological Principles

Full size image

Respect

The respect principle is highly relevant across all types of Smart City interventions that involve humans, e.g., that target human perception or behavior or that make use of human-generated data to undertake certain actions.

Challenges to Privacy and Autonomy. The European Commission’s rolling plan for ICT standardization regarding Smart Cities and Communities even explicitly concern the respect for individual autonomy and privacy by stating that ICT in Smart Cities should be used “with proper regard for security, both of individuals and their personal data, and use it as a driver for economic and social improvements” (European Commission - joinup, 2023).

Juvenile Ehwi et al. (2022) argue that there is need for a debate on Smart City data ethics and privacy even with legal guidance such as the European General Data Protection Regulation (EU GDPR) in place. Legislation may not only change over time but may also vary across countries and depending on the technology included. Furthermore, legislation may provide useful guidance, but sometimes lacks detail on how to meet the formulated requirements. An example of this are the Cookie banners often encountered when visiting a website that ask for the user’s privacy preferences. While intended for ensuring informed consent in line with EU GDPR, deceptive patterns applied to make users accept all cookies counterfeit its purpose (Utz et al., 2019). Hence, new regulations such as the ePrivacy Regulation are being drafted to keep up with and mediate new developments Footnote 1. This example illustrates that law and ethics are thus not equivalent. Following Floridi and Tadeo’s (2016) argument, compliance with legislation is deemed insufficient to determine what “good” moves could be. The need for a debate despite existing legal requirements is further emphasized by current empirical studies highlighting privacy as a concern for the citizens (Emami-Naeini et al., 2023).

Emami-Naeini et al. (2023) find that indeed, the conceptualization of privacy as a right to be free from unwanted intrusion can be in conflict with Smart City interventions that aim for control and automation. Caron et al. (2016) identify further challenges arising from the globality of the data collection, the mobility of sensors and their multi-disciplinary use.

An additional challenge is identified with regard to obtaining informed consent in data collection practices (Caron et al., 2016; Kitchin, 2016). While already challenging for literate, adult citizens, the challenge might be even more pronounced for vulnerable groups such as children or people with cognitive impairment who might not be able to read or comprehend privacy notices. Consider for example smart streetlamps that collect sensitive data of passersby, such as video footage, body temperature or gait. A privacy notice attached to the streetlamp might prove insufficient to inform all affected citizens about the data collection and give them the opportunity to opt out by avoiding a certain area. In addition, a relevant aspect to consider is the ongoing nature of providing consent: People might consent to data collection in one instance or for a certain type of data, but not for all types of data or forever. For the streetlamp example, a person might consent to having their body temperature collected but not the video footage. Thus, possibilities to dynamically adjust and restrict consent to specific applications should be considered.

Privacy-related Implications. Transparently informing about privacy and data practices is a first step towards acceptability of Smart City interventions. Previous projects already demonstrated that data being collected unawares is deemed unacceptable by citizens and can even lead to the abolition of projects (Figueroa, 2020; Stegall, 2020). An example is provided by the Sidewalk Toronto/Quayside project. A partnership of Alphabet’s subsidiary Sidewalk Labs and the public corporation Waterfront Toronto envisioned a smart neighborhood called Quayside enabled through driverless vehicles and innovative sensing technologies to track public activity (Artyushina, 2023; Woyke, 2018). After ongoing public debates and open questions with regard to the envisioned data collection practices and potential commercial interests, the project was canceled (Artyushina, 2023). However, the case contributed to legal and administrative reforms such as the Digital Infrastructure Plan, that requires city departments to prioritize open-source solutions when commissioning products from Smart City vendors (Artyushina, 2023).

Yet, besides mere information on the data collection, citizens should also be informed about the impact of data collection scenarios on individuals and the availability of controls (Emami-Naeini et al., 2023). In this context, Colnago et al. (2020) also discuss open challenges in Smart Home settings with regards to power dynamics that may also become relevant in the larger city context, e.g. between residents and visitors, parents and children, or privacy-active and privacy-passive partners.

However, the sheer number of interconnected and data-collecting devices in a Smart City may pose a challenge for the scalability of privacy notices and privacy settings. Requiring citizens to read, process, and handle numerous privacy notices individually might be overwhelming and lead to counterproductive effects such as anxiety and mental fatigue (Emami-Naeini et al., 2023). Previous research on privacy policies already demonstrated the impracticability of lengthy and complex privacy notices that are rarely considered by users (Obar & Oeldorf-Hirsch, 2020). Thus, support tools that aid with decision-making and implementing privacy preferences across devices need to be considered. As an example, Colnago et al. (2020) envision a privacy assistant for the Internet of Things (IoT) that helps users detect and manage the data collection practices of devices and sensors in their vicinity. An additional measure might be to limit notifications to data types perceived to be concerning, e.g. video footage (Emami-Naeini et al., 2023).

A complementary approach to informing citizens lies in reducing the amount of collected and shared sensitive information right from the start, such as.

  1. (a)

    through technical measures, such as by using low resolution cameras or counting people without storing images (Woyke, 2018).

  2. (b)

    through limiting collection and sharing practices, e.g., sharing data only with appropriate parties such as law enforcement (Emami-Naeini et al., 2023).

  3. (c)

    through involving citizens in the process of determining what level of data collection and sharing is deemed acceptable and beneficial, e.g., in participatory co-creation processes (Emami-Naeini et al., 2023).

Challenges to Respect for all People

The second aspect of the respect principle concerns respect for all people regardless of e.g., individual, or cultural differences, thus prescribing non-discriminatory practices. This aspect is also found not only applicable but also important in the context of ubiquitous data collection and invisible algorithmic decision processes that can reinforce existing inequality or bias (Cavazos et al., 2020; O’Neil, 2017). Bianchini and Avila (2014) for example describe the possibility of video surveillance systems to be trained to not only detect suspicious behaviors per se but to focus on individuals of specific ethnic groups that are supposedly more likely to be linked to criminal activities. These practices would lead to discriminatory and unequal treatment.

Example

The example of the smartphone application aimed at motivating people to save energy described in Sect. 3 touches the respect principle in several ways. The collection of information on personal energy use and housing situation and the integration with and potential sharing of personal information with the data of local energy providers may raise privacy concerns. Autonomy might be compromised if the intervention, i.e., the social norm nudge, exerts its influence unawares and is not transparent for the user.

Beneficence

Ensuring beneficence is challenged by the tension between measures targeted at the common good or society, and individual benefit or protection from individual harm, respectively. An example is posed by surveillance measures aimed at increasing public safety while potentially impacting personal privacy. Likewise, interventions informed by overall societal aims such as sustainability might not only lead to certain groups benefiting more but even some groups experiencing harm. As an example, imagine a financial incentive for using smart heating technology and tax increases for the remaining heating options at the same time. Then, people having the finances and the power to influence heating technology, e.g., homeowners, can benefit from the intervention. However, people who cannot afford the smart heating option or even influence the heating technology in their home, e.g., as tenants, experience a financial disadvantage. Certain interventions, such as political efforts to make citizens save energy, may even result in measures that are not beneficial for any group in the short term but directed at the future public good. While these governance-related challenges are not unique to the Smart City context, they are nonetheless relevant and without easy solutions.

Apart from objectively quantifiable beneficence, e.g., in financial terms, beneficence can be perceived differently depending on contextual or demographic factors, such as safety interventions being perceived as less useful in areas already perceived as safe (Emami-Naeini et al., 2023; van Heek et al., 2016). Another example is that people with more young children tended to rate the beneficence of IoT data collection scenarios as higher (Emami-Naeini et al., 2023). Asking for and including the perceptions and evaluations of different stakeholders is therefore essential for the reflection on beneficence. Potential measures to increase beneficence or perceived beneficence, respectively, thus include participatory measures in the design phase and evaluation measures in the implementation and operation phase to ensure that stakeholder perceptions influence Smart City interventions and their persistence or abolition.

Example

The example shows that while the intention is to serve the common good and to address global challenges such as energy consumption, beneficence cannot be uniformly assumed but still needs to be considered and evaluated. In the example, not only can homeowners with influence on energy use in their home and respective financial resources to install alternative heating solutions benefit more from the intervention as others such as tenants that cannot influence the installed heating system. But also, the intervention can even lead to a counterproductive increase in energy use for people already consuming little energy based on the descriptive social norm nudge used and a human tendency towards the average.

Justice

As envisioned by psychological ethical guidelines detailed in Sect. 4.3, the justice principle concerns access to and benefit from interventions for all persons (American Psychological Association, 2017). According to Ziosi et al. (2023) an important aspect of the ethical debate around Smart Cities is which benefits citizens will get and who will benefit from them. They outline that the input and active contribution of the citizens and other stakeholders are essential for a fair deployment of Smart City technologies. Thus, justice can be deemed an applicable as well as relevant ethical aspect.

Measures to achieve just access to and benefit from Smart City interventions and technologies mentioned by Ziosi et al. (2023) include a range of passive and active participatory approaches such as digital platforms on which citizens can vote on projects, citizens acting as sensors, or co-creating approaches.

Yet, challenges to this approach lie in their accessibility and the level of engagement (Ziosi et al., 2023). For example, technological or digital interventions such as hackathon-based forms of citizen engagement may only be appealing and accessible to tech-savvy citizens and exclude other groups from the process (Trencher, 2019). Furthermore, for more passive interventions, such as reporting problems or voting on only pre-defined projects, citizens are not involved in a creative design process. Thus, Trencher (2019) calls for more inclusive approaches that bring together both tech-savvy and less technical people in creative and problem-solving activities when developing Smart Cities.

Still, technological literacy is not only relevant for the creation of Smart City interventions but also with regards to who can benefit from them. If interventions are only accessible online or via using certain technologies, the already existing digital divide may be entrenched or even increase (Ziosi et al., 2023). People who lack digital literacy, access to the respective technologies, or even a sufficient internet connection may be treated “unjustly”, i.e., not having equal opportunities to access potential benefits. Accessibility as a relevant aspect of usability furthermore plays an important role for large groups of citizens, including children, elderly people, or people with disabilities.

Finally, discriminatory practices resulting from the technology and algorithms such as using data for profiling need to be mentioned again as they may also hinder access of certain groups to the benefits of Smart City interventions, e.g., when older people are marginalized by technological innovations (Sourbati & Behrendt, 2021; Ziosi et al., 2023).

Considering the justice principle in the Smart City context can take various forms from increasing awareness of potential biases to guidelines for the accessible and inclusive design of digital applications and additional interventions correcting for existing biases or compensating other limitations of existing interventions.

Example

The use of a smartphone application as described in the example also touches the justice principle as it might impose burdens to access and profit from the interventions for different groups of people, such as those with low technical affinity, elderly people having difficulties to handle the digital application or people with no Internet access. Therefore, consideration should be given to either decreasing these burdens or finding alternative ways to access the intervention, such as in a non-digital format.

Scientific integrity

Even though not all interventions might be initiated by scientists, but also politicians or those working in industry, for example, the principle of scientific integrity remains generally relevant for all Smart City interventions. Regardless of the type of interventions or their disciplinary focus, scientific standards and insights should be followed. In this regard, ethical psychological principles furthermore suggest being aware of one’s own competencies and their limits (The British Psychological Society, 2012), implying that interdisciplinary approaches are necessary for intervention design in complex Smart City environments. Depending on the intervention this can include considering:

  • psychological findings for interventions targeting human perception and behavior,

  • usability guidelines for accessible and intuitive interface design,

  • computer and data science principles for ensuring data quality, security, and privacy or.

  • mathematical models to ensure the quality of predictions or analysis.

One challenge to scientific integrity may be posed by the temptation to engage in “easy” Smart City interventions; Even private individuals who lack any expertise or awareness of scientific standards can, for example, quickly develop and launch a smartphone application. The application could have an unexpectedly wide reach and impact on the affected users which poses unanticipated ethical challenges. Another challenge is posed by the quickly developing, fragmented and technologically driven field which may lead to the design of new technologies and interventions based on their sheer possibility as compared to interventions being driven by user needs or considerations of potential down-sides (Oliveira & Campolargo, 2015; Zhao et al., 2021). Additionally, a challenge concerns the reduced level of control over Smart City interventions applied in the field and influenced by various unforeseeable factors such as social, environmental, or political aspects, as compared to classical laboratory trials that aim to limit external influences and keep differences between trials to a minimum to enhance internal validity (McGrath, 1981). And finally, Kitchin (2016) describes methodological challenges stemming from the shift from previously data-scarce analyses to big data analyses in the Smart City context.

A starting point is awareness for the interdisciplinary expertise required to design, implement, and evaluate a Smart City intervention. Again, an example is lent by data-driven nudges such as the smart streetlight project in the Netherlands that aims to reduce crime by calming light settings based on predicted aggressive behaviors (Hoogeveen et al., 2018). For the successful implementation of this intervention, technical expertise for the real-time data collection and analysis as well as psychological insights into human perception and behavior are necessary. In addition, the intervention might require criminological expertise for selecting suitable locations as well as legal expertise for handling the data collection and processing.

Example

In the described example, the city administration partners up with local researchers to design the smartphone application. Thus, one may assume that scholarly standards of the involved disciplines may be adhered to and that advising commissions often available within university contexts such as an ethics commission or a data protection officer may be consulted. Yet, as also a partnership with the energy provider is described, consideration should be given to ensuring that standards and regulations also apply to the energy provider and the data handled by it.

Social responsibility

Researchers are also increasingly calling for consideration of the social and long-term effects of Smart City interventions on vulnerable groups. An exemplary challenge described by Ziosi et al. (2023) is that Smart City developments on the one hand can involve tearing down neighborhoods to make space for new buildings while attracting especially high-skill professionals and entrepreneurs on the other hand. Both aspects contribute to the increase of housing prices thereby reducing affordable accommodation for certain groups.

Another aspect falling under the social responsibility principle is the sustainability aspect of the Smart City definition provided in Sect. 2. For example, Toli and Murtagh (2020) reviewed 43 smart city definitions regarding the inclusion of a sustainability aspect and found that almost all of the sustainability-oriented definitions also included social aspects of sustainability. It concerns ensuring that “the needs of present and future generations are met in the full respect of economic, social, environmental as well as cultural aspects and values of society (p.3)” (United Nations - United 4 Smart Sustainable Cities, 2021). Likewise, the standard ISO (2019) “Sustainable cities and communities — Indicators for smart cities” stresses the sustainability aspect and puts people at the center. Responsible use of resources and other sustainability efforts are ultimately supposed to benefit the people’s quality of life. The standard suggests that smart cities increase sustainability by enhancing citizen engagement, using collaborative leadership methods, and applying interdisciplinary approaches supported by modern technologies and data information. Thus, when designing and implementing Smart City interventions, the long-term consequences for future generations and their environment must also be considered. In that context, Ramirez Lopez and Grijalba Castro (2020) also consider the aspect of resilience, i.e. a city’s capacity to handle challenges resulting from changes and hazards such as natural disasters. Thus, challenges to sustainability are not only the optimization and fairness of resource use but also accounting for foreseen and perhaps unforeseen circumstances that can endanger the Smart City and its inhabitants.

An additional challenge not new to that aspect is the trade-off between the economic dimension and the social or environmental dimension, respectively (Ziosi et al., 2023). And while researchers agree that many Smart City interventions have the potential to increase sustainability overall, the development of required new technologies, their operation, and their recycling is resource- and energy-intensive (Ziosi et al., 2023).

Another challenge for considering social responsibility in Smart Cities stems from characteristics of the Smart City. These include difficulties in establishing or determining system borders, in analyzing isolated or long-term effects given quickly evolving technologies and complex interactions. Potential countermeasures include the integration of evaluation options and measures for adapting or ending the intervention right from the start should unanticipated side effects be detected. Furthermore, making use of existing long-term measurements collected in other contexts such as citizen statistics and regular surveys on various societal topics may provide initial indications for produced long-term effects.

Example

The intention to encourage energy saving aligns with the sustainability aspect included in the social responsibility aspect. However, the described example showed that unintended and counterproductive outcomes are possible so that these should be evaluated to adapt the intervention accordingly. In addition, social effects such as certain groups of people potentially benefiting more from the intervention as compared to others should be considered here.

Discussion and directions for future work

Selection of principles

This paper positioned psychological principles as a starting point for reflecting on ethical principles for all kinds of interventions in the interconnected and digitalized Smart City context that involve or affect humans. This choice was made as psychology as a discipline concerns human perception and behavior across various application areas. The principles are thus not limited to a certain technology such as robotics or certain types of interventions, such as medical ones. Yet, the author is aware that numerous other associations have established ethical principles in place that may also be relevant to the Smart City context. Most likely, many of these even bear some overlap with the proposed guidelines, such as the principles of biomedical ethics (Beauchamp & Childress, 2001). The four values Respect for Autonomy, Beneficence, Non-Maleficence and Justice show parallels to the psychological ethical principles described in this article. In addition, further ethical principles and legal requirements may become relevant or even enforced for specific application areas such as medical applications, artificial intelligence, or dual use applications. Dual use describes applications initially developed for good purpose that have the potential to also be easily misused for harmful secondary purposes, such as to be used as weapons (Forge, 2010).

To gain a more holistic and interdisciplinary perspective, it would thus be beneficial to compare and extend the presented psychological principles with those of other disciplines or associations. For example, Jobin et al. (2019) investigated and mapped eleven ethical guidelines and principles for “ethical AI” from existing guidelines across multiple organizations. As AI becomes increasingly relevant within Smart City technologies, these recommendations share a number of parallels with the psychologically- grounded ethics considerations presented here. The reason may be that in both cases the principles focus on the interaction with some kind of “intervention” in a broad sense. For example, among the eleven principles uncovered by Jobin et al. (2019) are Beneficence and Non-maleficence or Responsibility which echo the beneficence and social responsibility principles discussed above. In contrast, Freedom and Autonomy or Privacy which could be considered subdimensions of the respect principle are listed separately which might put more emphasis on each aspect.

As one example for existing AI guidelines the UNESCO recommendations (UNESCO, 2021) focus on Human rights and human dignity, Living in peaceful, just and interconnected societies, Ensuring diversity and inclusiveness and Environment and ecosystem flourishing. However, the values seem to put more emphasis on not only respecting but also ensuring diversity, as compared to the psychological principles. And whereas the psychological principles concern social responsibility, the value Environment and ecosystem flourishing seems to encompass a broader understanding of the human—environment interaction. Based on comparisons between different sets of relevant ethical principles and reflecting on their applicability, a broader and more holistic understanding of the ethical principles and values that should shape the future design of Smart City interventions can be reached.

Approaches to citizen participation

As outlined in Sect. 5, scholars call for citizen participation (Emami-Naeini et al., 2023; Trencher, 2019). Even though such a participatory approach may not directly solve ethical concerns or dilemmas, it is a valuable approach towards identifying ethical concerns, reflecting on them, and including the perspectives of those concerned in the process. The literature differentiates levels of participation ranging from non-participation to active participation where citizens have power, e.g. as co-creators or leaders (Cardullo & Kitchin, 2019). Passive, less engaging approaches range from informing citizens about planned interventions over listening to their opinion to having people vote on pre-defined projects as a “service-user” (Cowley et al., 2018; Trencher, 2019) or having them act as sensors (Ziosi et al., 2023). More active and inclusive approaches start where citizens actually have a say, e.g., as advisors or decision-takers. However, as compared to only voting on pre-defined options, active approaches at the upper end of the scale involve citizens as co-creators or proposers of new ideas ideally already in the need analysis or design phase.

However, according to Dieckmann (1998) one should also clarify the target group and the time point or planned duration of the involvement. For example, the target group may be citizens with a certain demography or specific groups such as elderly people, employees, or children. Furthermore, participants could be involved right from the start in the planning of an intervention, only when first ideas need to be decided upon, or only to evaluate an intervention. Participatory approaches can be one-off measures, take place at several time points, or regularly not necessarily tied to one specific intervention. These considerations should inform the selection of a suitable method, e.g., surveys or in-person co-creation workshops.

Ideally, the participation of citizens across the intervention lifecycle from need analysis to evaluation is envisioned. One example is provided by the project AktVisFootnote 2, short for “Action towards a joint vision”, that aimed to develop unused areas within small towns. First, citizens were involved via a survey and a workshop to identify user needs and quarters with development potential. They had the option to actively contribute ideas and be informed about the ongoing project by using an online tool. To increase awareness, public announcements, walking tours, and chalk spray information on the sidewalks were used. Second, follow-up workshops were held with people from the quarters identified in the first step to collect and discuss concrete project ideas. As an alternative to the online tool, an analog version was also provided in the second workshop. The third step focused on the people directly concerned by the project ideas, e.g., the owners of a respective building or space, and included advisory meetings with them on possibilities for implementing a project.

Besides concrete projects, there are several examples for platforms or initiatives that foster inclusion of citizens and exchange between research and the public. Examples are the eu-citizen.science initiativeFootnote 3, the “Citizens create knowledge” platform in GermanyFootnote 4 or the Citizen Science CenterFootnote 5 in Switzerland that foster participatory research approaches.

Missing aspects and directions for future work

The Smart City literature so far has mainly focused on the technological aspects of making a city “smart”. However, others, among them Emami-Naeini et al. (2023), Meijer and Bolívar (2016) and Ziosi et al. (2023) emphasize the human aspects of Smart Cities, such as “smart citizens”, the active engagement with and contribution of citizens, as well as their perceptions. They view humans not only as a relevant part of the Smart City definition but also as a critical contributor to the success of Smart Cities and the acceptance of related interventions. As such, greater emphasis should be on the human element in Smart City research and practice. The reflection on applying psychological ethical principles to the Smart City context can be seen as a first step in that direction.

In line with that, Sholla et al. (2017) propose to dedicate a layer within the Smart City architecture to socio-cultural and ethical aspects to shift the previous focus on technology more towards the people within Smart Cities. The aim of that layer is to explicitly establish limits and to inform ethics aware decision-making. Yet, future work should consider an even broader range of human aspects, such as how to best leverage human creativity to contribute to Smart City design, research on human perceptions of beneficence and justice, or research extending the work on ethical principles to values or norms.

Conclusion

This article reflected on the applicability of principles for ethical psychological research and practice to the Smart City context. Smart Cities are viewed as socio-technical systems so that many interventions involve humans, e.g., through the collection of sensitive data, and thus require the consideration of ethical principles. Smart Cities are further characterized by a high degree of complexity and interconnectedness, that lead to emergent system outcomes, cascading effects, and an unclear reach of interventions.

Therefore, ensuring Respect, Beneficence, Justice, Scientific Integrity, and Social Responsibility poses challenges in the Smart City context as compared to previous small-scale data collection in laboratories or sampled surveys. Reviewing each principle, concrete challenges were identified, such as the difficulty to obtain informed consent for ubiquitous data collection and to ensure all citizens can engage in and equally benefit from Smart City interventions regardless of individual differences. Additionally, challenges lie in preventing unanticipated side effects in an interconnected and uncontrollable setting and in foreseeing long-term effects given the complex interplay of technological, social, and environmental aspects.

Whereas previous work has often focused on the technological aspects of Smart Cities, this and related research call for a stronger focus on the engagement and interaction with the citizens who actually make a city a Smart City. The ethical aspects and their implications presented here provide a starting point but should be extended by a broader understanding of ethical principles and values. Initial suggestions can be derived from comparisons with relevant ethical frameworks of other associations.