What started off as rather tame research on integrating care into retail robots, ended up as a creative and political call to upend increasingly automated food systems. Ley argues that robotics are the next step in a long history of separation from the human and non-human life involved in getting food on the table. The dissertation itself documents the author's unravelling as she comes to a concluding ethical vision: a messy and inefficient life of interconnection, led by the senses.

The dissertation draws together myriad academic fields, from philosophy of technology, phenomenology, robot-ethics, feminist theory, ethics of care, science and technology studies, and decolonial practices. Woven around the traditional academic chapters are poems, stories, creative prose, post-its, and photographs. By stretching into the creative and intuitive realms, the dissertation envisions what philosophy of technology might look like.
... Read more

In the Netherlands, there has been a growing exposure of discriminatory practices across various domains. Notable instances such as the child benefit allowance scandal (referred to as the "toeslagen affaire") and concerns surrounding student allowance (DUO) have captured public attention. These problems are not confined to public domains; instances of discrimination are also being found within the financial lending markets.

While existing research offers critical discussions, practical analyses, and technical solutions regarding discrimination, there is a lack of well-defined design principles specifically addressing discrimination challenges in financial lending in the Netherlands. This study aims to bridge the gap by providing a concrete set of design principles that architects can utilize to effectively design non-discriminatory financial lending systems, while also recognizing the potential for policymakers to draw inspiration from these principles and develop universal guidelines or other policies to address this issue more broadly.

This study combined the Design Science Research approach and Principle Based Design to derive design principles for a socio-technical topic like non-discrimination. The study started with identifying the factors contributing to financial discrimination in the lending system through desk research and 13 semi-structured interviews with a wide range of scientific experts and industry professionals. To make the concept of discrimination more applicable in the context of financial lending, the study defined specific dimensions which are differential treatment, differential outcome, unequal access, and lack of transparency. This study then identified and categorized various factors that can contribute to one or more dimensions of discrimination into six domains: regulatory, human interaction, user experience, business, data, and technical. These domains help identify the specific areas in which discrimination can arise within the lending context and how different factors may overlap across multiple domains. This also highlighted the complexity of discrimination, demonstrating that discrimination is not limited to one domain, and addressing a single factor can have cascading effects, either reducing or exacerbating discrimination across other domains.

The study proceeded to identify the challenges encountered in realizing a non-discriminatory financial lending system in the Netherlands. This was accomplished through a combination of semi-structured interviews and literature review, which led to the identification of 13
challenges. Design principles were then developed that could address those challenges in collaboration with seven business and IT architects. These principles were also evaluated in two workshops involving another group of architects. Based on the feedback received, efforts were made to improve the structure and consistency of the principles, resulting in a final set of 11 principles. To enhance clarity, the principles were rearranged and grouped into four categories: culture, governance, process, and models.

The study acknowledges that while the identified design principles can contribute to reducing discrimination in the financial lending system, they do not fully resolve the problem. This is because many discriminatory issues are deeply rooted in society. However, the principles are valuable additions to existing practices and tools, providing a framework for organizations to improve their lending practices. In addition, the involvement of regulatory authorities is essential to address discrimination in the lending system. Without regulatory incentives or oversight, organizations are not motivated to proactively implement such principles or other initiatives aimed at promoting non-discriminatory lending practices. Moreover, the design principles have the potential to be applied in other areas, such as governmental financial allowances, extending their impact beyond the financial lending system.
... Read more

These days, autonomous driving has become a popular topic in different fields and showed its unavoidable developing tendency. With the increasing application of autonomous vehicles, a mixed period of autonomous vehicles (AVs) and conventional vehicles (CVs) is inevitable. Therefore, some traffic management was proposed. Exclusive lanes for AVs are a new way to improve the mixed road environment's traffic capacity. However, the effect of exclusive AV lanes was rarely studied. Therefore, the decision-makers could not clearly understand this new policy's impact. Besides, the existing research content did not provide a detailed quantitative analysis of this new policy, which is very important for decision-makers. Thus, this thesis focused on the exclusive AV lanes and proposed the following research question: Under which conditions are exclusive autonomous vehicle lanes societally beneficial? ... Read more

In the past decade, rapid shifting and evolving technological systems that take trust as part of the design objective (call such systems “trust-inviting systems”) have incredibly transformed the way we interact with others. Think of reputation-based platforms mediating interactions between strangers and other digital services provided by institutions, as well as blockchain systems, which work to shape, respectively, our trust relations with individuals, institutions, and technologies. However, the ways these systems use to facilitate trust are not always justifiable and can lead to negative moral and social consequences if certain parts of the systems are shown to be flawed. By reflecting on several present cases of trust-inviting systems that are experiencing great tension of trust, this thesis argues that trust-inviting systems essentially attempt to interpret, translate, and ultimately institutionalize the idea of trustworthiness in given contexts. However, the ways that trust-inviting systems are using to institutionalize the characteristics of trustworthy persons, institutions, and technologies should not be accepted without scrutiny. For each case analysed here, a discrepancy between the intention to improve trust and trustworthiness and the means that are adopted to facilitate them is shown. Such cleavage is argued to be primarily caused by flawed understanding of the trust concepts and the resulting ill-suited design choices, as well as problems emerged from the implementation process. These issues are proposed to be ameliorated by a recalibration of the understanding of the trust concepts, which has the potential for remedying shortcomings of the current design and development of the systems with forward-looking strategies taking into account a wide range of societal needs, values, and technical properties. In a word, it is argued that trust-inviting digital systems should be designed, developed, and deployed in ways that are aligned with the essence of the trust relation in context, in order to achieve proper trust and trustworthy systems. As such, the pitfalls identified in each case are able to be used as perspectives contributing to building affordances that foster warranted trust and foreclosing affordances that would undermine warranted trust.
... Read more

The use of machine learning systems has great potential to better predict probabilities of default for credit underwriting. Despite this advantage, herewith there exists the substantial risk of discrimination. Moreover, machine learning models with the highest prediction-accuracy are often the least explicable (i.e. explainable). Nonetheless, explicability is needed to create accountability of automated credit decisions by machine learning systems. Furthermore, there exists a regulatory need for explicability of machine learning systems in the General Data Protection Regulation (GDPR) and the Consumer Credit Directive (CCD). Besides that, an ethical- and societal need exists for explicability. Within the exploration of literature, it becomes clear that research lacks on how to move from a high-level principle like explicability, towards a prospective assessment of a machine learning use case on this principle, it lacks a multi-disciplinary perspective, and it misses an assessment framework that can guide decision-makers within machine learning use cases, aligned with a multi-organizational development lifecycle. This research aims to design a prospective pragmatic assessment framework that can guide decision-makers, within machine learning applications in European credit underwriting cases from the point of view of explicability. To accomplish this, the Design Science Research Methodology (DSRM), complemented with the Value Sensitive Design (VSD) approach, is utilized. To this end, the Explicability Assessment Framework (EAF) was developed. This framework is adapted to the context- and explanation characteristics of the case, and aligns with the CRISP-DM development lifecycle. It was found in two case studies that the framework helps with the decision-making whether a machine learning system is sufficiently explicable or not. Lastly, a wide range of future research areas is identified that needs attention: empirical validation and expansion of the framework, the relevance for automated explanation creation, the scalability to other context and a large amount of explanations, and the practical perspective regarding adoption in the industry. ... Read more

The European Commission introduced PSD2 with the purpose to promote innovation by increasing competition and participation of non-banks in the payments industry. Amongst others, PSD2 requires banks to allow a way for third-party providers to have direct access to a current account’s data if authorized by the customer. This forces banks to open up their databases which calls into life a particular stream of questions about how technologies should be designed in order to preserve privacy. Therefore this research was dedicated to designing a framework of general principles for the inclusion of privacy in innovation within the financial industry, in particular for the incumbent bank.

The Value Sensitive Design approach was utilized in order to provide this framework through a qualitative explorative case study. In the conceptualization stage of VSD, privacy was defined through a literature review. For the empirical investigation, data was collected through semi-structured interviews with eleven experts from three of the four types of stakeholders. Values that were identified in literature were qualitatively validated by scoring their importance relative to each other. For the technical investigation, the most important values were translated into design requirements by utilizing the value hierarchy method. The application of VSD to innovation in finance is the theoretical contribution of this research.

Privacy was conceptualized as complex and interrelated with other values. Values related to privacy in the case of PSD2 were also identified in several literature streams. Through a qualitative validation, the final set of values to be reflected by technology designed after PSD2 was determined which shows that privacy is the value of greatest importance to be included in financial innovation. Also, based on the insights experts brought forward during the interviews, five ethical challenges that come with the implementation of PSD2 were identified. The applicability of the value hierarchy method was demonstrated by translating the value privacy into six design requirements. These six design requirements addressed the end-user’s ability to make an informed decision for consent, ex-ante privacy assessment through standardized licensing procedures and a shared license database, the possibility for end-users to manage information disclosed through APIs and the limitation of information referring to transaction’s counterpart that did not consent. ... Read more

Artificial intelligence is a technology trend that potentially far exceeds the possibilities of all current state technologies, in a sense that these technologies have self-learning abilities. Organizations that manage to adopt and diffuse these technologies may effectively reach a competitive edge over their competitors faster. However, for large organizations such as traditional banks the question arises how to facilitate this transformation to effectively adopt AI technologies. Despite the fact that AI technologies are still in a relatively premature stage of development, traditional banks in the Netherlands are currently exploring its potential. In this exploratory study we interviewed 19 experts in the field of AI and Banking to identify all relevant challenges that arise when traditional banks adopt AI technologies. Additionally, we explored the importance of top management tech-executive roles as a possible solution to address certain challenges. As a result we have found that traditional banks have no common view or approach on how to adopt AI technologies enterprise-wide effectively. Therefore, the appointment of a new Chief AI Officer role may prove to be a valuable solution to address some of these challenges. However, there is still a mixed view on what this roles’ position, responsibilities and other managerial characteristics should be. ... Read more

Automated driving systems (ADS) can improve traffic safety, improve accessibility and reduce environmental impact (Shladover, 2016). On the contrary, on May 7th 2016, a fatal accident with a Tesla on autopilot in U.S. Florida was a harsh reminder that the technology is still in its testing phase (Greenemeier, 2016). In complex technical systems like ADS technical failure may occur, which forms a serious threat to human well-being. Moreover, studies have shown that citizens are very concerned about deliberate misuse of ADS (Kyriakidis, Happee, & De Winter, 2015), e.g. people purposely abusing ADS to cause damage or even hurt someone. Therefore, according to many experts, the implementation of ADS does not only entail technical issues but also normative issues. To bridge the gap between technical- and normative issues, responsible innovation can be applied (Santoni de Sio, 2016). Friedman et. al. (2006) propose value-sensitive design to achieve responsible innovation via technical-, empirical- and conceptual research. As following from responsible innovation and value-sensitive design and in particular its focus on empirical research as one of its necessary elements, in this research an effort is made to use empirical research methods to provide more insights in normative issues of automated driving systems (ADS). The focus lies on social acceptance, particularly with respect to traffic safety. Also accessibility, environmental impact and heterogeneity among citizens will be analysed. Social acceptance is defined as “a person's assent to the reality of a situation, recognizing a process or condition (often a negative or uncomfortable situation) without attempting to change it, protest, or exit” (Fish, 2014, page 1). The following questions will be answered in this research: What is the social acceptance of automated driving systems from the perspective of safety, accessibility and environmental impact and what is the corresponding heterogeneity? 1. What percentage of citizens thinks automated driving systems are socially accepted? 2. How is the social acceptance influenced by safety, accessibility and environmental impact? 3. Are traffic fatalities caused by automated vehicles valued differently than current traffic fatalities? 4. Is there heterogeneity in the social acceptance among citizens? A survey is chosen as research method since it is a relatively inexpensive, flexible method to achieve extensive information about characteristics of a population. After an extensive theoretical analysis, seven attributes were identified that possibly influence the social acceptance: level of automation, road exemption, travel time, emissions, human error fatalities, technical failure fatalities and deliberate misuse fatalities. After the experiment was fine-tuned by a pilot study, it was held among a representative sample of 510 Dutch adults during the spring of 2017. The respondents had to state if they were in favour or against ADS for each of the twelve hypothetical futures that were presented to them. In these hypothetical futures, the attributes were systematically varied and described as a change to the current situation. Using a MNL RUM model, the results show that 63% of all citizens prefer ADS over the current system. It is therefore concluded that citizens have a high social acceptance and thus are rather positive towards ADS. Also, citizens prefer a system where human drivers are still in control and can intervene in case necessary. Next, it is concluded that the social acceptance is mostly influenced by fatalities caused by automated vehicles (AVs), while travel time is the least important attribute. However, the differences in influence of the attributes were not substantial. Safety, accessibility and environmental impact are all important for the social acceptance. Nevertheless, technical failure fatalities weigh as much as 4 human error fatalities. For deliberate misuse fatalities this is a factor 5.5. Although these relations coincide with literature, the magnitude is larger than expected. It implies that ADS have to be very safe in order to reach social acceptance. Since AVs are still ‘learning’ how to drive, this might cause problems for current and future experiments. A latent class choice model is estimated to answer the final research question. Results show that large heterogeneity exists among citizens in the social acceptance. Citizens can be segmented into three classes (% of citizens): automated driving enthusiasts (32%), central mass (52%) and risk-averse class (16%). Contradictory to average citizens, automated driving enthusiasts prefer high automation levels. Even so, they still weigh fatalities caused by AVs as much as 3 human error fatalities on average. The central mass shows similar results to the results of the MNL model estimated on the full sample. The risk-averse class has a strong dislike for fatalities caused by AVs. This class weighs technical failure fatalities (deliberate misuse fatalities) as much as 5.5 (10) human error fatalities. In conclusion, primarily two discrepancies are identified that are critical for the implementation of ADS: 1) High social acceptance versus strong dislike for fatalities caused by AVs; 2) Citizens who are enthusiastic about ADS versus citizens who are risk-averse. They lead to the following recommendations: The social acceptance for ADS is high, so it is recommended for policy makers to have a positive and active approach towards ADS. By conducting experiments for professional users, safety risks can be minimalized while a learning curve is ensured. Also technology producers and policy makers should intensify research into cooperate driving. According to experts, ADS and cooperate driving are inseparable (Shladover, 2016), but globally the research into cooperate driving is lacking (Roland Berger, 2017). Since cooperate driving can lead to an increased risk of deliberate misuse, it is deemed critical for the implementation of ADS. Next, policy makers and especially the RDW should review the licensing of AVs. Currently, hardly any restrictions are in place for the licensing of AVs, which can cause dangerous situations on public roads. Finally, information campaigns can help to make citizens aware of the risks and benefits of ADS. ... Read more