Former DSI Fellows

Prof. Dr. Naomi Baron

Dr. Naomi S. Baron was a DSI Fellow between October 16th-29th, 2019. During those two weeks, she participated in the following activities:

Lectures and Meetings
Anchoring her work were a URPP Language and Space Colloquium (“Wayfaring on the Ground and Onscreen: What to Abandon, What to Embrace”) and a public DSI lecture (“Does Medium Matter for Learning? Reading in Print, Onscreen, and with Audio”). Other activities included meeting with colleagues to discuss shared research interests (Andreas Jucker, Christa Dürscheid, Volker Dellwo, Sara Fabrikant, and DSI Fellow Lonneke van der Plas) and with doctoral students in the URPP Language and Space program. 

Collaboration on Research Project Planning
Together with Elisabeth Stark, Simone Ueberwasser, and Marina Bondi (from the University of Modena and Reggio Emilia), we began mapping out a project to explore communication choices and opportunities for older citizens using mobile devices. Building on a model of citizen science, the initiative is designed to improve the lives of senior citizens in the Zurich area by identifying current usage patterns and needs, and then creating training programs that expand usage possibilities.

Progress on New Book
The rest of my time was devoted to the book I am writing, "How We Read Now: Effective Strategies for Print, Digital, and Audio". Enjoying the welcoming workspace of the RAF building during its final days of housing the DSI, I inched ahead both in the writing and in conceptualizing later chapters.

For over a decade, I have been examining the role that technology plays in how we read. My 2015 book "Words Onscreen: The Fate of Reading in a Digital World" examined the history of reading, reading formats and media, and results of my own empirical study of undergraduate practices and preferences regarding reading in print versus on a digital device. That book focused on identifying the pros and cons of reading on each platform, drawing on research up through 2014. But much has changed since then, both in the research world and in educational practices. There are a host of new studies (especially in Europe and Israel) documenting cognitive differences in the ways students approach the two media. At the same time, in both lower and higher education, we are seeing increased use of audio (or audio-visual) materials in lieu of text, as well as a march to replace paper-based standardized testing with digital-only.

The intended audience for How We Read Now is educators: teachers, school administrators, policy makers, and parents. The book summarizes contemporary research findings, which become the basis for suggesting practical strategies to maximize student learning when reading in print, digitally, or with audio.
I am grateful to the DSI for the opportunity to work on this book in such a congenial atmosphere.

Website: Prof. Dr. Naomi Baron

Dr. George Bruseker

Dr. George Bruseker was DSI Fellow from April to November 2019.


My DSI fellowship was structured in order to support work on the project “Building a Semantic Foundation for Architectural Data Integration”. The project aim is to work on the development of semantic modelling for the areas of art and architecture in order to build reference models for representing advanced CH information in a way that would support sustainable and repeatable advanced analytic research on different questions in this domain. While scholars increasingly adopt digital methods and formats for documenting analytic information regarding the history and evolution of art and architecture in its material and conceptual aspects, insufficient focus is placed on the development of common conceptual models for representing this basic research data. Such common representations, however, are necessary not only to support research in the present but to facilitate its reuse in the future and the possibility of its integration with other domains of inquiry. My research on this topic in the context of this fellowship concentrated on covering the question of how to represent and query creative processes in art and architecture. As part of a larger discourse on Digital Society, this research plays into the question of the re-democratization of digital assets and making accessible research knowledge to society and social actors in a reusable and comprehensible manner.

My research period began in May and was structured to take place largely in Zurich over the summer of 2019 (June and July) with follow up until December of 2019. Research in conceptual modelling requires access to primary documentation systems and close collaboration with researchers in the domain to understand their data documentation practice and its relation to their research questions. Being based at the DSI enabled a close collaboration with University of Zurich’s SARI fellowship holder, Anais Guillem, and the team of SARI in order to collaborate over the case study of the Heinz Isler archives ( A case study in architecture was of exceptional use to this research as the architectural creation process highlights the interdisciplinary, multi-phase and multi-actor nature of creative processes in architecture in particular but also arts in general by extension. Moreover, a stay in Zurich itself was highly enriching and stimulating to my research, enabling visits to the rich museum collections in Switzerland as well as the highly active and enriching exhibition programmes.

This period of study allowed for the co-development of a semantic data model extending the CIDOC Conceptual Reference Model and the harmonized submodel, FRBRoo, an ontology for bibliographic data. The semantic data model which resulted from this research, Creative Processes Representation (CPR), is a general model providing ontological classes and properties for describing the evolution of an artistic oeuvre over time: from programme, through design, execution planning and execution. 


The academic interaction and funding support that the DSI provided supported continuous work on the questions of creating reusable models for representing CH data and allowed me to create and deliver lectures on this topic to multiple audiences during the course of this work. These included:

“Semantic Reference Data Models”, w/ Nicola Carboni,
Bits and Bites Session, Fourth Swiss Congress for Art History 
Mendrisio, Switzerland

“Some thoughts on speech acts and historical modelling” 
Historical content metadata workshop, Apollonis Programme
Heraklion, Greece

“Modeling the models: how semantic representation can transform our use of architectural archives?”, w/ Anais Guillem 
University of Zurich, SARI / ETH Zurich 
Zurich, Switzerland

“Formal Ontologies, the Semantic Web and CIDOC CRM: what are these things and what is their relevance to digital archaeology?”
College Year in Athens
Athens, Greece

“Democratizing Data in Digital Society: re-claiming a critical relation to information through semantics, with a case study in researching creative process” [tentative title]
University of Zurich, Digital Society Initiative (DSI) 
Zurich, Switzerland

Outreach and Networking

In addition to the work undertaken in Switzerland, I was able to participate in several international fora for the use of semantic data, particularly committees and groups aiming to build the tools to support access to knowledge through semantic data. In these contexts, I was able to bring up the on-going research in Zurich, gaining feedback and contributing to international research and development on the topic. These outreach programmes included:

CIDOC CRM SIG (11-14 June, BNF, Paris)
Linked Conservation Data (London, Sep 12-13, 2019)
LinkedArt (Victoria and Albert, London; Oxford University, Oxford, Oct 1-4, 2019)
DONIPAT : École thématique «Données Interopérables pour le Patrimoine» (Aussois, France, Oct 14-18, 2019)
CIDOC CRM SIG (22-25 Oct, ICS-FORTH, Crete)
Arches Resources Modelling Working Group (Getty Foundation, Nov 19-21, 2019)

In 2020, I will aim to publish the outcomes of the research in international journals.

Website: Dr. George Bruseker

Dr. Moritz Büchi

Dr. Moritz Büchi was a DSI fellow from August 2019 to July 2020 and worked on the topic of digital well-being. 

The fellowship enabled a time of focused theoretical work on the connections between digital media uses and subjective well-being for which the DSI provided an intellectually and physically ideal environment. During my time at the DSI I was able to complete a manuscript whose conceptualization had started in 2018. A related empirical focus on the concept of perceived digital overuse was published [] open access in Social Media + Society. The motivating question for this project overall was how people can improve or maintain their well-being both thanks to and despite the constant availability of digital media and resulted in “blueprint” for explanatory theories. Digital media is increasingly intertwined with everyday life, therefore, their potential impact on overall well-being has also grown, often leading to overly generalized and dramatic claims, such as "smartphones are ruining a generation". The resultant framework is briefly described in this video [] and highlights three interdependencies for well-being in the digital society that demonstrate the implausibility of “easy answers”: First, individuals’ digital practices depend on the opportunities and constraints afforded by their social surroundings and technological developments. Second, different manifestations of individuals’ digital practices lead to often co-occurring concrete harms and benefits. And third, the balance between and cumulation of concrete harms and benefits will affect overall well-being. There is nothing inherently beneficial or harmful in digital media per se, but they can and do impact people’s subjective well-being, opening a normative discussion about adequate governance towards promoting positive impacts and keeping negative impacts in check.

Beyond the fellowship itself, the Challenge Area Communication [] continues to provide a platform for collaborative efforts and feedback. Inspired by the discussions around different concepts, a couple of colleagues and I have launched a digital well-being reading group and a project on disconnection strategies.


Website: Dr. Mortiz Büchi  


Dr. Linda Di Geronimo

Dr. Linda Di Geronimo was DSI Fellow from May to November 2019 and was working on the following topic:

Dark Patterns (DPs) are user interfaces that trick the users into doing something they did not mean to do. Sneaking unwanted items into the basket, adding users to sometimes costly subscriptions, or using double negatives on checkboxes are only some of the many DPs that can be found online. While there are occasions where designers do not introduce these patterns intentionally but rather from inexperience and a lack of user testing, they often come from clear malicious intents.

DPs do not only drastically lower the user experience, but they can also lead users to over-share personal information. Although the effect of DPs on our society is clearly relevant, there has been little effort in the research community on how to best tackle this problem. Moreover, while the concept of Dark Patterns is not new, a rigorous survey on their extent and their impact on the user has yet to be carried out. To bridge this gap, I initiated research on the topic.

Thanks to the support of the Digital Society Initiative, I was able to classify Dark Patterns in trending mobile applications (N=240). We found that among the tested apps, 95% contain at least one or more Dark Patterns. The classification was performed by first recording ten minutes example usage of each app (for a total of 40 hours recordings). Then each UI and its interaction were evaluated by two researchers. The researchers followed a previously developed taxonomy and kept note of each malicious design and its category. We will soon provide the classification together with the final dataset on this website:

Furthermore, together with Ethix, we are building a list of guidelines for a more ethical digital user experience. The goal of this phase is to develop a label to evaluate web and mobile applications through an ethical lens.

Website: Dr. Linda Di Geronimo

Prof. Dr. Martin Dusinberre

As a DSI Fellow in autumn semester 2019, Martin Dusinberre was working on the final stages of a major digital history project funded by the Swiss National Science Foundation.

Lives in Transit is a tool both to narrate and to teach global history in a digitized age. It offers historians a model for presenting their research in the form of serious gaming; and the game itself simulates the real-life experience of research for BA and MA student-players. Unlike conventional academic formats—the essay, the chapter or the monograph—the game allows the writer to explore nonlinear storytelling and to create parallel and multi-directional narratives. Through gamification, Lives in Transit invites players to experience the genesis of a piece of global history research and offers them innovative ways of chronicling and mapping this creative process. At the same time, it encourages them to reflect on the nature of doing research in a digitized world. With its strong focus on the present-day research environment of academic work, it points to the unevenly distributed access to research funding and even online sources—problems which reflect the bigger debates over access in the “digital society.” Lives in Transit is an open source project and the result of a long collaboration between Professor Dusinberre, the S3IT department of the University of Zurich and private partners. The prototype was developed and tested in 2017; the game itself is being tested in autumn 2019 and will be launched in spring 2020. 

Website: Prof. Dr. Martin Dusinberre

Dr. Sarah Ebling

Dr. Sarah Ebling was DSI Fellow from January to September 2019 and worked on the following topic:

Towards Automatic Text Simplification for German

While digital transformation has brought about positive change for many people, it has also created additional barriers. For example, the fact that digitalization has led to more and more information being provided in written form is an obstacle for people with reduced reading ability. Easy language can give these people access to information and communication. Such a language variety is characterized by reduced lexical and syntactical complexity, the addition of explanations for difficult words, and a clearly structured layout.

The automatic production of light language is already reasonably well researched for languages such as English, Spanish, Portuguese etc. For German, on the other hand, there is almost no work. The aim of the Fellowship "Towards a Scalable, Multimodal Automatic Text Simplification System for German" was therefore to create the conditions for the development of an automatic text simplification system for German.

The concrete output of the Fellowship are two parallel corpora that will serve as data collections for automatic text simplification via neuronal machine translation (Battisti and Ebling, 2019). The first experiments in this direction are already underway. In addition, a doctoral student at the Institute of Computational Linguistics began his research on monolingual sentence annotation, which provides a basis for the usability of parallel documents for neuronal machine translation.

Since one of the two corpora set up within the framework of the Fellowship consists of news reports, an exchange with Prof. Dr. Anne Scherer, Prof. Dr. Thomas Friemel and Prof. Dr. Chat Wacharamanotham was established within the framework of a project dealing, among other things, with a digital infrastructure for news texts.

In the current autumn semester 2019 I am also conducting a seminar "Automatic Text Simplification" (together with Prof. Dr. Martin Volk). In the spring semester 2020 I offer a seminar "Language Technology in the Context of Artificial Intelligence", inspired by the exchange at the DSI.

The Fellowship also provided an opportunity for networking with other Fellows and members of the DSI network. Together with Prof. Dr. Eva Weber-Guskar, I conducted a "Science Café" at Scientifica 2019 entitled "Can a computer system think?" With Prof. Dr. Adrian Hehl (Vetsuisse Faculty) an application for a teaching credit project was submitted.

Website: Dr. Sarah Ebling

Prof. Dr. Thomas Friemel

Prof. Dr. Thomas Friemel has been working at the DSI in FS18 and HS18 on 3 projects:

Communicative Challenges in Digital Societies

Both public and private communications have changed significantly as a result of digitalisation. The lecture series of the Digital Society Initiative in cooperation with the IKMZ addressed from different scientific perspectives the most important opportunities and challenges which arise for society as a result of these changes. Among other things, it discussed what influence new forms of online propaganda and fake news have on political debates, how online platforms can be legally regulated, what strategies media companies can use to assert themselves in changing market conditions, how children and young people can learn to use the Internet positively and to what extent robots can help people with disabilities to participate in social exchange. However, the influence of digitalisation on scientific research was also discussed. Digital traces of Internet use and electronic devices such as smartphones provide researchers with new insights into people's communication and relationship behaviour. At the same time, these data sources are associated with numerous problems, ranging from ethical and legal aspects to the question of what digital traces stand for. All lectures of the DSI ring lecture were filmed and are available on the DSI website. 

Protection of Children Online

Against the background of the increasingly developing digital landscape and the associated risks for children and young people, 60 experts from all over the world exchanged views on questions such as "what new online risks arise in a digital landscape" or "what critical digital skills are needed to raise risk awareness among children and young people". The workshop discussed the foundations on which the new OECD recommendations should be based. 

Digital Health Communication

Both communication and health were identified as relevant challenge areas within the framework of the DSI. This indicates that the UZH has both the necessary critical mass of scientists and a strategic interest in playing a leading role in these fields. The declared goal of DSI is not only to stimulate research and interdisciplinary cooperation at the UZH, but also to communicate the resulting expertise to society and to support the transfer of practice. The development of a CAS in Digital Health Communication has exactly this goal. On the one hand, the identification and specification of relevant subject areas and the expertise existing in them at the UZH will strengthen the networking of researchers. On the other hand, the transfer of know-how will contribute to the development of the digital society and underline the leading role of the UZH.

Website: Prof. Dr. Thomas Friemel

Dr. Alfred Früh

Dr. Alfred Früh was a DSI Fellow in 2019. During this time, he focused on how patent law is challenged by technological developments. The focus was on the question of how to deal with "inventions" made by machines. More and more machines are developing technical solutions almost autonomously. This could be considered "inventive", would patent law not require human inventors. In view of this contradiction, the division of tasks between man and machine (which is completely non-transparent from the outside) becomes decisive and raises the question to what criteria patent law should adhere when granting patents. The study suggests extending inventorship to undertakings and to banish traditional elements of inventor personality right from patent law. This would ultimately create more transparency about who actually is (or was) the inventor of a given invention. 
This topic is part of a broader habilitation project that examines how law deals and should deal with transparency in the digital economy. 

Website: Dr. Alfred Früh

Prof. Dr. Fabrizio Gilardi

During the DSI Fellowship I held in the Spring and Fall semesters 2019, I could invest significant time and effort in setting up the Digital Democracy Lab (, which is a key component of the DSI Challenge Area Democracy. The Digital Democracy Lab is now fully functional. It consists of a research infrastructure capable of collecting, pre-processing, storing, and analyzing a large amount of data with a high degree of automation. The infrastructure is hosted on UZH’s computing and storage infrastructure, which ensures data encryption and protection. This database includes, for instance, millions of social media posts related to Swiss federal elections, 10 million newspaper articles from 86 Swiss news organizations, over 40,000 press releases from Swiss parties, government actors, and interest groups, as well as complete transcripts of roughly 500 editions of Switzerland’s most important political talk shows. These datasets are updated continuously and can be extended depending on researchers’ needs, also for political and media sources outside Switzerland. To establish international cooperations, we have issued a call for non-resident fellowships ( Furthermore, we issued a call for papers for a DSI Challenge Area workshop scheduled for May 20, 2020, which, unfortunately, we now had to postpone (new date to be defined) due to COVID-19. A second, international workshop is planned for November 5-6, 2020. In sum, the DSI Fellowship played an important role in kick-starting the Digital Democracy Lab. I am grateful to the DSI for its support.

Website: Prof. Dr. Fabrizio Gilardi

Dr. Jamie L. Gloor

From September to December 2020, Dr. Jamie Gloor was employed part-time (60%) at the DSI. During this time, she published and presented several works related to digitalization, particularly as applied to the workplace. For example, together with Professor Dr. Carolin Strobl and Dr. Rudolf Debelak (UZH Psychology), she published an article in Inside IT outlining concrete recommendations for leaders, computer scientists, and organizations to reduce algorithm anxiety. In addition, she published a thought piece on how soft skills such as humor may power robot leadership in The European Business Review together with Professor Dr. Lauren Howe (from UZH business at the future of work center), Professor Dr. David De Cremer (founder and director of the Centre on AI Technology for Humankind at the National University of Singapore) and Professor Dr. Kai Chi (Sam) Yam (National University of Singapore Business School); elements of Dr. Gloor's theorizing on soft skills and robot leadership were also published in a widely read technical report on "The Future of Work". During her DSI fellowship, Dr. Gloor presented, "The future of work (post?) COVID-19: New leadership challenges" for the Exeter Festival of Social Science. Just before starting her fellowship, Dr. Gloor also presented a methods talk at the DSI on "Research in times of corona-virus: Challenges & creative solutions." In addition, Dr. Gloor also assisted with organizing the DSI conference, "Trust in Autonomous Machines," and planned to present at this conference and at a DSI fellows event, but unfortunately could not do either due to illness. Finally, Dr. Gloor met and enjoyed valuable exchanges with current/former DSI fellows/scholars, including Professor Dr. Aniko Hannak, Drs. Markus Christen, Markus Kneer, Ulrich Leicht-Deobald, Michele Loi, and Eleonora Vigano (to name a few).

Website: Dr. Jamie L. Gloor

Prof. Dr. Miguel Gonçalves Meira e Cruz

Prof. Dr. Miguel Gonçalves Meira e Cruz was visiting Zurich and the DSI in August 2018 to work with Dr. med. Dominik Ettlin.

DSI fellowships for translating the web-based interdisciplinary symptom evalua-tion1 (WISE) tool into Portuguese, thus enabling collection of transcultural “big data” for improving the understanding of risk factors for chronic orofacial pain 

Patient reported outcome measures gain increasing importance in various medical areas. We re-cently designed and constructed a web-based interdisciplinary symptom evaluation (WISE) for patients suffering from orofacial pain and temporomandibular disorders1. For worldwide applica-bility and to avoid copyright infringements, open source software tools and free validated ques-tionnaires available in multiple languages were used. Highly secure data storage limits access strict-ly to those who use the tool for collecting, storing, and evaluating their data. By combining a symptom- burden checklist with in-depth questionnaires serving as case-finding instruments, an algorithm was developed that assists in clarifying case complexity and need for targeted expert evaluation. This novel modular approach provides a personalized, response-tailored instrument for the time- and cost-effective collection of symptom-burden focused quantitative data. The tool includes body drawing options and instructional videos. It is applicable for biopsychosocial evalu-ation in a variety of clinical settings and offers direct feedback by a case report summary. The tool thus enables personalized medicine, facilitates interprofessional education and collaboration, and allows for multicenter patient-reported outcomes research.

Almost 1,000 patients of the interdisciplinary pain unit of the Center of Dental Medicine of the University of Zurich already completed the WISE prior to their initial consultation. Data analysis for research projects is in process and manuscripts are submitted to scientific journals or in prepara-tion. In order to enable the translation of the WISE for use in other countries, the DSI Board of Directors had financed two fellows in August 2018, Prof. Antônio Sérgio Guimarães (University Leopoldo Mandic, Campinas, Brazil) and Prof. Miguel Gonçalves Meira e Cruz (University of Lis-bon, Portugal). The two were joined by Prof. Kanokporn Bhalang (Chulalongkorn University in Bangkok, Thailand) who had been able to raise other funds for her stay in Zurich. The team of three worked intensively for 7 days on the translation of the software into their respective lan-guages and the work was successfully completed within the given timeframe. After work, family members joined for leisure time (photo).

Team orofacial pain
Left to right: S. Guimaraes, Ray (husband of N. Bhalang), Maria (daughter of M. Meira e Cruz), Ruth (mother of M. Meira e Cruz),  M. Meira e Cruz, N. Bhalang, D. Ettlin 

Website: Prof. Dr. Miguel Gonçalves Meira e Cruz

Dr. Andrew M. Guess

Despite the unexpected obstacles to scholarly exchange presented by the coronavirus, my stay at the DSI allowed me to make important progress on projects related to the Initiative’s mission.

Specifically, my time at DSI provided me with the space and feedback to complete two papers related to digital literacy and online misinformation. The first, with Kevin Munger, attempts to clarify the concept of digital literacy and explore how best to measure it. We also talk about practical challenges for social scientists interested in studying how digital literacy relates to outcomes that we care about. One of the survey measures we look at was originally developed by DSI affiliate Eszter Hargittai, so it was a privilege to be able to speak with her in person about the project. I was also fortunate to receive helpful feedback from a DSI Brown Bag Lunch and the Fellows Exchange Day.

Relatedly, while at DSI I finished a project on how digital media literacy interventions on social media can help people identify online “fake news.” That paper was published in the Proceedings of the National Academy of Sciences recently and generated a lot of discussion of our findings. In particular, we discovered that “tips” on how to spot fake news can measurably improve discernment between false and mainstream news posts in both the United States and India.

Finally, I was able to launch several new projects that I am continuing to work on. For example, I’ve begun a study on exposure to online (mis)information related to Covid-19 and the effectiveness of official corrections on people’s misperceptions related to the outbreak. I look forward to updating DSI on this project as it develops!

Website: Dr. Andrew M. Guess

Prof. Dr. Antônio Sérgio Guimarães

Prof. Dr. Antônio Sérgio Guimarães was a DSI Fellow at the UZH in August 2018 and worked with Dr. med. Dominik Ettlin.

DSI fellowships for translating the web-based interdisciplinary symptom evalua-tion1 (WISE) tool into Portuguese, thus enabling collection of transcultural “big data” for improving the understanding of risk factors for chronic orofacial pain 

Patient reported outcome measures gain increasing importance in various medical areas. We re-cently designed and constructed a web-based interdisciplinary symptom evaluation (WISE) for patients suffering from orofacial pain and temporomandibular disorders1. For worldwide applica-bility and to avoid copyright infringements, open source software tools and free validated ques-tionnaires available in multiple languages were used. Highly secure data storage limits access strict-ly to those who use the tool for collecting, storing, and evaluating their data. By combining a symptom- burden checklist with in-depth questionnaires serving as case-finding instruments, an algorithm was developed that assists in clarifying case complexity and need for targeted expert evaluation. This novel modular approach provides a personalized, response-tailored instrument for the time- and cost-effective collection of symptom-burden focused quantitative data. The tool includes body drawing options and instructional videos. It is applicable for biopsychosocial evalu-ation in a variety of clinical settings and offers direct feedback by a case report summary. The tool thus enables personalized medicine, facilitates interprofessional education and collaboration, and allows for multicenter patient-reported outcomes research.

Almost 1,000 patients of the interdisciplinary pain unit of the Center of Dental Medicine of the University of Zurich already completed the WISE prior to their initial consultation. Data analysis for research projects is in process and manuscripts are submitted to scientific journals or in prepara-tion. In order to enable the translation of the WISE for use in other countries, the DSI Board of Directors had financed two fellows in August 2018, Prof. Antônio Sérgio Guimarães (University Leopoldo Mandic, Campinas, Brazil) and Prof. Miguel Gonçalves Meira e Cruz (University of Lis-bon, Portugal). The two were joined by Prof. Kanokporn Bhalang (Chulalongkorn University in Bangkok, Thailand) who had been able to raise other funds for her stay in Zurich. The team of three worked intensively for 7 days on the translation of the software into their respective lan-guages and the work was successfully completed within the given timeframe. After work, family members joined for leisure time (photo).

Team orofacial pain
Left to right: S. Guimaraes, Ray (husband of N. Bhalang), Maria (daughter of M. Meira e Cruz), Ruth (mother of M. Meira e Cruz),  M. Meira e Cruz, N. Bhalang, D. Ettlin 

Website: Prof. Dr. Antônio Sérgio Guimarães

Prof. Dr. Christian Hauser

Prof. Dr. Christian Hauser worked at the DSI from June to December 2019 on the following topic:

In communication and media studies, framing describes the process by means of which themes and events are classified into patterns of interpretation. According to Entman (1993), frames fulfil two functions. On the one hand, they serve to select certain aspects of reality and on the other hand, these aspects are to be linked in communication with certain problem definitions, interpretations, moral evaluations or instructions for action.
For some years now, the term "big data" has stood for the drastic changes brought about by the increasing penetration of digital technologies in society. A precise definition of the term does not yet exist. However, it makes it clear that nowadays large amounts of very different data can be easily captured, stored and analysed. Not only computers, smartphones and portable sensors, but also cars, household appliances, machines and buildings routinely generate large amounts of data thanks to embedded information and communication technology (ICT).
The project explores which frames are used in Swiss and US newspapers in the debate on big data. To this end, the opportunities and risks associated with "big data" and presented in a communicative way in the newspaper debates in Switzerland and the USA will be identified. On the one hand, the differences that exist in the newspaper discourses of the two countries are to be worked out and on the other hand, how the frames to "Big Data" have changed in recent years is to be examined. Furthermore, it will be examined which companies and industries shape the reporting on "Big Data" and how the entrepreneurial use of "Big Data" is ethically evaluated.

Website: Prof. Dr. Christian Hauser

Prof. Dr. Christoph Heitz

Prof. Dr. Christoph Heitz was DSI Fellow from April 2019 to March 2020 and worked together with ethicists of the DSI on the project 

Algorithmic fairness and ethics of commercial data use 

Algorithmic fairness is about algorithms that, for example, calculate individualized insurance premiums, design personalized information supply, control predictive policing, or implement applicant selection using artificial intelligence. Such algorithms are being used more and more today, and they thus intervene noticeably in the social fabric. 

The question of the DSI project was: How can algorithms be constructed that meet the ethical expectations of fairness, justice and non-discrimination of customers, employees and society? Answering this question requires a combination of ethical discourse (What exactly do we mean when we speak of fairness or justice?) and technology (How to construct fair decision algorithms? How can fairness-by-design be realized technically?). 

As a result of the DSI Fellowship, a joint project with the DSI members Markus Christen, Michele Loi and Aniko Hannak ("Socially acceptable AI and fairness trade-offs in predictive analytics") was launched in the NRP77 programme of the SNSF.  

In the somewhat more general context of data ethics, the project is concerned with the development of ethical guidelines for the commercial use of data in the context of data-based services. Under the leadership of Christoph Heitz, Markus Christen and Michele Loi, a project group of scientists and company representatives developed an "ethical code for commercial data use", the final version of which will be published in autumn 2020. A first version is available here. These activities took place within the framework of the Swiss Alliance for Data-Intensive Services and were financed by Innosuisse.

Video on Youtube

Website: Prof. Dr. Christoph Heitz

Dr. Markus Kneer

Having previously been employed at the DSI, I could hit the ground running and benefit to the full from its extraordinary academic environment. Existing collaborations with Markus Christen, Avi Bernstein, Suzanne Tolmeijer and former fellows Mike Stuart and Eva Weber-Guskar were continued. I also got to know a number of new scholars, in particular Aniko Hannak, with whom I started a new project.

During the one-year fellowship, we finished the TA-Swiss expert study “Wenn Algorithmen an unserer Stelle entscheiden: die Herausforderungen der künstlichen Intelligenz” under the leadership of Markus Christen and published it as a book. A joint survey paper on moral algorithms with S. Tolmeijer, C. Sarasua, M. Christen and A. Bernstein was accepted for publication at ACM Computing Surveys. A joint NFP project proposal entitled Meaningful Human Control in Security Systems with T.Burri,  M. Christen and F. Fleuret received funding. I finished a paper entitled Can a robot lie? as well as Guilty artificial minds (with M. Stuart), both of which are currently under review. With I.Hannikainen I conducted the largest study on moral dilemmata in Corona triage contexts to date (under review), which sparked considerable interest from the media.  With a team from the DSI, armasuisse and Kobold games, we also made progress on a large-scale study using computer simulations to explore trust in AI, a topic on which we also organized the DSI Conference Trust in AI held in November 2020 and on which I spoke at the Swiss Digital Days. Finally, Markus Christen and I are currently finishing our draft on lethal autonomous weapon systems and responsibility gaps.

Website: Dr. Markus Kneer

Prof. Dr. Thomas Lampoltshammer

Ass.-Prof. Dr. Thomas Lampoltshammer conducted research at the DSI from February to December 2020 on the following project:

How innovation networks can strengthen companies and regions.

The effects of digitalization are currently more noticeable than ever. While this paradigm shift was initially understood as a technical problem, it is now already a central component of strategic governance concepts and thus core to the new value proposition (Li et al., 2017, Grover et al., 2018). This impact on value creation has also been recognized by the European Commission. Here, the latter speaks of approximately €415 billion per year in economic growth, in support of employment, competition, investment and innovation in the EU, underpinned by concepts such as the Digital Single Market and related regulations such as the Single Digital Gateway Regulation (SDGR). In order to effectively leverage the repositioned value proposition, there is a need for a correspondingly well-established flow of information and thus relationships between the company and the respective business partners. The ability to leverage appropriate networks by the company - especially for SMEs, as they have access to comparatively fewer resources - therefore becomes a critical component (Battistella et al., 2017; Parida et al., 2017). By expanding network capacities, companies are able to respond promptly to market changes and thus to dynamic demands on resources and, moreover, to identify new lines of business more quickly as well as to make them usable (Acosta et al., 2018). It is precisely the strategic management of internal and external networks (and information flows) that allow SMEs to increase their performance, reduce overhead in processes, thereby saving costs, and ultimately share knowledge and participate in shared knowledge (Lin & Lin, 2016) - keyword: open innovation (Chesbrough, 2003).

The project proposal addresses the tension between 'exploitation' and 'exploration' and the resulting performance implications (Acosta et al., 2018) in companies. A complementary approach could combine both advantages here (Kristal et al., 2010), but quite a few companies fail at this hurdle (Solís-Molina et al., 2018), also due to the necessary and, depending on the framework conditions, very heterogeneous requirements for structure and resources (Gonzalez et al., 2018); this is especially true for SMEs, as they usually have very limited resources (Junni et al., 2013). Here, digital networking can help SMEs to improve their performance, especially also towards a 'Digital Platform Capability' (Cenamor et al., 2019). The project proposal supports this approach by using state-of-the-art methods of Natural Language Processing (NLP) and Social Network Analysis (SNA). By means of apportionment to the concept of stigmergy, it becomes possible to harness the online self-representation of companies and to formulate and validate a conceptual innovation network along the companies' thematic orientations as well as their competencies.

During the last year, the DSI Fellowship offered me the opportunity to further elaborate the theoretical foundations of the project. Among other things, initial experience was gained with regard to the processing of unstructured, heterogeneous data in the context of the stigmergy approach. This included a comparison of different methodologies and algorithms with regard to data processing and analysis. Furthermore, an initial concept for data fusion using ontologies could be designed and first walking tests using text-based networks could be performed. Unfortunately, due to the COVID-19 pandemic starting in 2020, there were massive restrictions, not only with regard to the stays on site, but also with regard to the initial project plan due to the exceptional situation and its effects at my home institution. Nevertheless, due to the progress described above, the state of Lower Austria was persuaded to follow up on the results of the fellowship with a project in early 2021 and to advance alternative and complementary research strands. The Fellowship has also given me the opportunity to further develop my cooperation with my DSI mentor Prof. Dr. Uwe Serdült. Another highlight of the Fellowship was the opportunity to exchange ideas with other Fellows and staff of the DSI. I would especially like to mention my meeting with Dr. Aurelia Tamò-Larrieux, and I would be happy to maintain and further develop this contact.

Website: Prof. Dr. Thomas Lampoltshammer

Dr. Juliane Lischka

Dr. Juliane Lischka was a DSI Fellow during FS18 and HS18 and worked on the topic "About Black Sheep and Sheeple - Deviant Agents, Counterknowledge, and the Good Information Society".

Thought Provoking or Bullshit? Credibility Cues of Conspiracy Theories

In an era of epistemic instability (Harambam, 2017), conspiracy theories are a common topic in online debates (Wood & Douglas, 2015). Conspiracy theories are easily distributable and accessible on platforms such as YouTube and have a long online shelf live (Del Vicario et al., 2016). 

This project aimed at developing a credibility cue model for conspiracy theory videos. Credibility of online information largely depends on source, message, author, and receiver characteristics (Metzger & Flanagin, 2015). For videos, production value is found to be a relevant credibility dimension (Cummins & Chambers, 2011). 

Results of this study show that regarding conspiracy theory credibility cues, respondents largely refer to these credibility cues, but also extend them with two effect-related aspects: (1) effects on personal user behavior and (2) effects on others. Regarding the first aspect, respondents argue that their intent to follow up on a conspiracy theory in personal communication or finding it thought-provoking affect their conspiracy theory credibility rating. Regarding the second, the perception of a conspiracy theory as popular and sticky is relevant for their credibility evaluation. However, results based on an experimental survey design reveal that when watching a conspiracy theory video, merely host reliability and informational value play a role for the credibility rating. That is, respondents largely evaluate conspiracy theories based on journalism-like accuracy standards. 

Moreover, host reliability and informational value are related to behavioral intentions of respondents, such as sharing the video with a friend. Thus, accurate appearing conspiracy theories will be distributed more broadly, which stretches their online shelf lives. A higher credibility assessment of conspiracy theory videos also corresponds to the intention of being more skeptical about the government and media, indicating a decrease of trust in societal institutions.

Website: Dr. Juliane Lischka

Dr. Julian Mausbach

Dynamic electronic consent and de-anonymization

Digitisation when including people and health data in research

The first area of the project is devoted to the question of whether the self-determined (intended) involvement in research can be made more favourable by digitalising the consent process. In terms of content, the main focus of the study is on making the process more dynamic. With respect to formal aspects, the focus is placed on the question of whether the process as a whole can be designed electronically.

The second area of the project approaches the involvement in research from the opposite direction, so to speak. It is dedicated to unwanted inclusion and does not focus on design possibilities, but rather on design needs due to digitalisation. Specifically, the question is whether the possible de-anonymisation of research data by digitalisation leads to a re-identified person being able to derive rights from it or to obligations arising for researchers as a result.

Website: Dr. Julian Mausbach

Article in inside-IT: Einwilligung in Forschung ist keine Einbahn­strasse!

Prof. Dr. Matthias Mehl

Prof. Dr. Matthias Mehl will be a guest at the DSI and the UFSP in July 2018:

The main project I pursued within my DSI fellowship was to help identify ways in which the ambient audio samples collected via the uTrail can be converted into meaningful indicators of daily social activity. We recently published a paper in the journal Psychological Science demonstrating that the simple metric of “percentage of time talking” is a robust behavioral marker of wellbeing (Milek et al., 2018). As a first step, then, it would be important to develop ways to automatically and accurately extract voice activity information from the ambient audio. Other interesting candidate indicators could be emotional expressions (e.g., laughing, sighing, crying) captured by the ambient audio recordings as well as prevalent daily activities such as watching TV, cooking and eating, commuting, etc. Interestingly, this task, that is the task that is currently most pressing with respect to analyzing data collected with the uTrail, happens to also be a highly timely task for and a rapidly advancing research area in the field of computer science. 

The field of behavioral and speech signal processing from naturalistic audio recordings is in the process of becoming a vibrant research area. This year, a group of researchers around Dan Ellis, a former Professor at Columbia University who has since transitioned to Google, published a seminal handbook that establishes the Computational Analysis of Sound Scenes and Events (Virtanen, Plumbley, & Ellis, 2018) as an active research area with great potential for significant scientific contributions in computer science and electrical engineering. The purpose of this field is to elevate audio signal processing to the level of sophistication of other “sibling” areas such as visual signal processing (e.g., face and object detection). Currently, the field is struggling to identify realistic data sources that are inherently important and provide the kind of “high noise” data that one encounters in everyday life. One recent attempt, Google AudioSet (Gemmeke, Ellis, et al., 2017), has focused on sampled snippets of YouTube clips which are realistic in nature but, arguably, capture only a limited, thin slice of people’s daily social lives. What the field is currently lacking is (1) a data base of naturalistic, real-world audio events that covers the full spectrum of lived daily life and that can be used to train computational models for the classification of a range of behaviorally important sound events (e.g., talking, laughing, sighing, eating) and (2) a data base of such audio events that is annotated for “ground truth” variables. 

With the data that is currently being collected via the uTrail device as part of a large ongoing project on healthy aging at the UFSP (200 participants wear the device for 4 weeks), we have access to an at-scale data base (ten thousands of sound bites) of real-world ambient audio clips sampled representatively from the full range of participants’ daily lives, morning to night, Monday through Sunday. And within the existing long-term collaboration between Dr. Martin and myself, we have access to a very large, “ground-truth” annotated data base of audio data based on the close to 20 years of research that Prof. Mehl has conducted with the Electronically Activated Recorder (EAR) method (Mehl et al., 2001; Mehl 2017). During my month as DSI fellow, I helped lay the groundwork for a collaboration that can take the computational analysis of everyday sound and speech events to the next scientific level. 

Website: Prof. Dr. Matthias Mehl

Dr. Gianluca Miscione

Dr. Gianluca Miscione was a DSI fellow from December 2018 until May 2019:

I have been DSI fellow from December 2018 till May 2019. This fellowship overlapped with the second half of my Sabbatical year, which I spent mainly at the University of Zurich. Being part of the DSI allowed me to extend the research I have been developing in collaboration with University College Dublin (my main affiliation), the Information Management Research Group(IMRG) at UZH, and the University of Muenster among others. 

My current research has been focused on blockchain and organizations (governance more precisely). Thus, a concrete step in the direction of relating to the DSI focal area on ‘Democracy’ has been the development of a literature review on ‘Blockchain and Democracy’. This document, together with the extensive collection of publicationson whose selection it is based, has been made available to all DSI members and also published online through Zotero, a free of charge reference management systems. This work has also been accepted for inclusion in the Blockchain Research Network Library, the most comprehensive collection of academic papers on all aspects of research on blockchain. The rationale of choosing free distribution has been to help any researcher who is interested in this emerging research topic, and ideally to attract some as possible collaborators. I would even suggest this mode of dissemination for other outcomes of the DSI’s activities.

Another part of my fellowship has been dedicated to networking within the DSI and across organizations. Regarding the former, the DSI fellows meetings have been the perfect format to go deep enough into others’ work and socialize. The book club (organized by prof Hargittai) was always very stimulating and led to this publication (Hargittai E., Miscione G., (2019) “Watch Me Play: Twitch and the Rise of Game Live Streaming” on the International Journal of Communication). Also the Breakfast of Ideas has been a useful series of meetings, which I have actively supported, for early stage researchers. Ideally, this may evolve into a more consolidated interdisciplinary venue for PhDs and Postdocs. 

Regarding networking across organizations, I was glad to introduce some members of the IMRG to related activities of the DSI. Hopefully this will lead to other fellowships and other more formal collaborations in the future. Beyond academia, members of the incubator Trust Square(and ProCivisespecially) and Crypto Valley have been made aware of the DSI and potential collaborations.

Last but not least, broader dissemination of ideas has been pursued by publishing ‘Der Blockchain-Hype – braucht es "Ockhams Rasiermesser"?’ [in German] for the Digital Society Initiative Insights series on Inside-IT, and giving interviews, one in the podcast The Baseline about “Currencies as infrastructures”.

I see many possible developments of those initial activities. The most immediate ones can about collaborations for joint applications for funding for international research projects. 

Website: Dr. Gianluca Miscione

Prof. Dr. Jean-Henry Morin

From September to December 2018 Prof. Dr. Jean-Henry Morin was a guest at the DSI and was working on the following topics:

Swiss Digital Transition and 2019 Federal Elections
In the context of the 2015 Federal Elections we joined forces with GovFaces to organise a public debate on the Swiss Digital Agenda trying to stimulate engagement on the issue among the parliamentarians, the political parties and the civil society ( While the debate did provide some interesting contributions and outcome, it failed to achieve the goal of including the issue in the political debate for the Federal Elections. Being one year away of the next Federal Elections in 2019 and given the still lagging public policy leadership on the issue that paved the way for the DigitalSwitzerland initiative (formerly Digital Zurich 2025) essentially driven by the economy, we think it it could be a very good opportunity to design a new initiative to counterbalance the economy-only driven approach and jointly work on a new initiative towards that goal in the context of the UNIGE-UZH DSI initiative. The State of Geneva with whom we are in close contact and collaboration through GenèveLab, the growing civictech area and the UZH contacts with the federal level are interesting supporting elements towards such a goal.

The Centre Universitaire d’Informatique (CUI) at UNIGE has been working for several years towards creating an academic FabLab. Its goal is to introduce the now common fabrication and DIY approaches originating the the work at MIT (Neil Gershenfeld), into the university to leverage the “build to think” and “build to learn” approaches based on design thinking, rapid prototyping and user centred design. Such approaches hold tremendous potential towards transforming the way we teach, learn and do research. Moreover they offer interesting opportunities for joint work between academia and the civil society. Recently the FacLab project at UNIGE received interest from the Division de l’Information Scientifique (DIS), the Bureau de la Stratégie Numérique de l’UNIGE (BSN), the Division du Système et des Technologies de l’Information et de la Communication (DiSTIC) and several external organisations such as Lift Conference, Genève Lab of the State of Geneva, etc. During the last Open Geneva festival in April 2018 we organised a hackathon to co-design the FacLab where participants from UZH joined (Prof. Florent Thouvenin). This led to several actionable outcomes which we will leverage starting this Fall 2018 in Geneva. We think it it could be a very good opportunity to jointly work on this project trying to build a sister FacLab at UZH in the context of the UNIGE-UZH DSI initiative.

Digital Responsibility
Based on research carried out over many years in the area of Digital Rights Management, particularly in the context of enterprise security, Governance Risk and Compliance (GRC) and Information Protection and Control (IPC), the issue of Digital Responsibility emerged as an important issue in the context of the Digital Transition and the responsible design of sustainable systems and services. The whole issue of cybersecurity and data protection is clearly in the need for profound changes in how the future of our digital society is to be shaped for the future (Morin 2014). 
Based on this, recent discussions took place with the Fondation Ethos in order to design a Digital Responsibility Criteria for their evaluations. This work will start in the Fall of 2018. In parallel, initial discussions with Johan Rochel
of Ethix (Lab für Innovationsethik) in Zurich. 
In this area, we think joining forces in the context of the UNIGE-UZH DSI initiative could accelerate and strengthen the impact in this important area towards a sustainable digital society.

Website: Prof. Dr. Jean-Henry Morin

Dr. Maël Pégny

From February 24 to April 3, 2020 Dr. Maël Pégny was a guest at the DSI and was working on the following topics:

My fellowship at the DSI was of course deeply impacted by the coronavirus outbreak. I did not have much time to enjoy the comfortable venue, the various events and the remarkable service provided by the administrative team. Roughly half of my stay in Zurich was spent working from my apartment via e-mail and Skype conferences with my two partners in Switzerland, Michele Loi (UZH) and Andrea Ferrario (Computer Science, ETH). I was not even able to have a physical meeting with Michele Loi, my main partner in this project, as he was unable to come back from Milan after the epidemic started there.

The initial targets of my stay in Zurich were to act as a consultant for projects on which Dr. Michele Loi and Dr. Markus Christen were working, and to write with Dr. Loi on a French-Swiss, ANR-SNF grant application on the epistemological and ethical stakes of Machine Learning opacity. Those objectives had to be scaled down because of the considerable disorganization created by the pandemic: I actually never had the opportunity to have a formal meeting with Dr. Christen. As a consequence, we decided to focus on our main objective, i.e. the redaction of the ANR-FNS proposal. I am happy to say that despite considerable adversity, we were able to submit in due time an application. 

Let me give a synthetic presentation of this project, and its interest for the DSI. Some of the most recent advances in Machine Learning are notorious both for their performance, e.g. their predictive power, and their opacity, i.e. our lack of understanding of their inner workings. This raises considerable issues from a technical perspective, as those pieces of software are hard to analyze, debug and generalize. They also raise issues from a social perspective, which can be summed up with the following questions: when is it legitimate to use an artefact we do not fully understand? As the industrial impact of Machine Learning continues to grow, this question is bound to have considerable impact on ethical, political and legal reflections. This was the original impetus for Dr. Loi and I to write a project gathering philosophers of science, who would try to understand the nature of Machine Learning opacity and the means to circumvent it, and moral and political philosophers, who would try to articulate the social consequences of such an opacity and define the legitimate use of opaque AIs. Such a project is not only timely: it answers an urgent need, as the philosophical reflection is lagging behind the fast technological changes, and its growing impact on society. 

The interdisciplinarity of our proposal, which was at the core of our methodology from the beginning, was only deepened by further work and ultimately lead to an opening towards legal scholars and computer scientists. We have joined force with Danièle Bourcier, a renowned founding figure of Computer Science & Law in France, to articulate the consequences of our ethical reflections in a white paper dedicated to law makers and legal scholars. We have also included the participation of scholars active in the field of statistics and Machine Learning. This came from the realization of a particular complementarity between philosophy and Machine Learning. There is a very active subfield of AI dedicated to the conception of tools to resolve or at least circumvent Machine Learning opacity, called eXplainable Artificial Intelligence. This field tries to produce explanations of the behavior of opaque systems. In the current situation however, there is on the one hand a lack of conceptual understanding of what a good explanation should be among computer scientists. On the other hand, there is a general lack of awareness of this technical work among philosophers, who have spent considerable effort to clarify what the desiderata of a honest explanation should be. Those two communities thus need each other, and we have decided to bring them together through a project of automated explainer system for health care, which will try to see if the philosophical conceptions of explanation can be implemented in computer systems and solve real-life problems of Machine Learning practitioners. This idea has also lead to increase the international nature of our project, as we have included as external partners scholars from Milano, who will explain their needs in facing Machine Learning Opacity in their daily practice. Finally, we have been joined by a computer scientist from the ETH Zürich, Andrea Ferrario, who will help us with our project of automated explainer. Our commitment to hands-on experience with computer science has led us to an original move for a philosophy project, which is the inclusion of a four-year computer science postdoc on our explainer system, who will travel between Paris and Zürich to implement the deep relations between the two teams and the two disciplines. 

Let me explain what the positive impacts of this proposal would be for the DSI. This should not only be an opportunity to take position on a trending subject, on which the Zürich canton authorities have been very active. It would contribute to the creation of an active research network between three countries on a new and promising interdisciplinary nexus: to the extent of our knowledge, we are the only project who brings the collaboration between philosophy and Machine Learning to this level of intimacy. This is a world first, which would earn a place at the cutting edge of the field for the DSI. The proposal takes also into account the importance of linguistic diversity for explanation, and our trilingual team will pay extra-attention to this issue, which is so important for the Confederation, in the development of our explainer system. Finally, the extreme relevance of our topic to current social evolutions is bound to open a venue between the most advanced academic research and public debates, be it through our legal work or otherwise. 

Website: Dr. Maël Pégny

PD Dr. Birte Platow

PD Dr. Birte Platow is a DSI Fellow from January to December 2019 and will deal with the following topic: 

Anthropomorphic Transferences as Constitutive of the Encounter between Man and Artificial Intelligence

How do individuals perceive themselves in dealing with artificial intelligence, and how do they behave as a result? This is the central question of my Fellow Project "Anthropomorphic Transferences as Constitutive of the Encounter between Man and Artificial Intelligence", which I am allowed to realize at the DSI in interdisciplinary exchange under optimal conditions.

The reason for my research is the observation that in the encounter between man and artificial intelligence the traditional frame of reference of anthropological descriptions of man changes implicitly and unconsciously, that the self-images and the corresponding behaviour of man are determined by new points of reference. To the extent that the individual is increasingly represented digitally, man and AI become complementary dimensions of a single and usually functionally determined reference system.

On the basis of a qualitative-empirical study, I would like to describe ideal-typical encounter situations between humans and Al as well as the patterns and strategies of self-perception and perception of Al that become recognizable in these situations. I interpret the findings and further questions arising from this study as a theologian against the background of Christian anthropology in order to contribute to the question of what ethics we need in a future in which the significance and scope of AI will be even greater than today.  

Website: PD Dr. Birte Platow

Dr. Mike Stuart

Mike Stuart was a DSI Fellow between February 2018 and January 2020, undertaking work toward the following project:

Project: How to Hold Machines Responsible for their Actions

Currently, there is no way to place legal blame on autonomous artificial agents (AAAs). But it is only a matter of time before they reach levels of autonomy approaching animal or human levels, at which point we may want to hold them (at least partially) responsible for their actions. For an AAA to be legally responsible, it must have a mens rea or “knowledge of purpose”; it must know in advance of acting what the consequences of its actions will be. But the precise nature of this epistemological state is still quite vague. For example, do we really require knowledge (in the philosophers’ sense), or would a state that comes in degrees (like understanding or awareness) be preferable? My project uses survey based methods to determine what the relevant epistemological state is, and draws on existing work in epistemology to identify what other features AAAs must possess in order to stand in that epistemological state.


I began by using online surveys to identify which mental states could be classified as mens rea, and whether laypeople naturally have a bias against the idea of robot responsibility. It seems that no such bias exists: what matters is that the agent responsible for harm was (at least) 60% certain that what they were doing would lead to harm. Nevertheless, I found that people were less willing to attribute beliefs to a robot or machine than they were to a human. I am now working on whether it is possible for machines to have occurrent or non-occurrent beliefs. I am also focusing on the asymmetry between our willingness to praise machines for good work and unwillingness to blame machines for harm they cause. I am also focusing on expert intuitions (instead of lay intuitions). My next project will concern the possibility of robot imagination, as well as “explainable AI.”

While at the DSI, I presented my work in a lecture entitled “Guilty Artificial Minds.” I also met with colleagues to discuss potential research collaborations (Markus Kneer and Eva Weber-Guskar), and received very valuable feedback from my other colleagues. I appreciated my time at the DSI very much.


Website: Dr. Mike Stuart

Dr. Aurelia Tamò-Larrieux

Dr. Aurelia Tamò-Larrieux is DSI Fellow from January to December 2020. Her research project is:

Decision-Making by Machines: Regulatory Strategies to Address the Unwanted Side-Effects of Automated Decision-Making

In January 2020 I started out with the planned mapping of side-effects of automated decision-making. The screening of the literature, together with interesting discussions I had at the DSI among others during the fellow exchanges, helped me to determine the scope of my first article. In that article I provide a taxonomy of the side-effects that arise from automatic decision-making and describe how legislation (with a focus on data protection law) may provide redress mechanisms to addressing these unwanted effects. This article has been accepted in a leading law journal with only minor revisions pending (the publication date will likely be spring 2021). Building upon this article, I have narrowed my research scope in three further contributions. Two of those contributions will be part of the HEC Symposium on Law and AI in 2021 that is planned to be held in Paris. While the first article focuses on the specific issue of discrimination, the second paper to be presented at the Symposium on Law and AI looks at the effect of automated decision-making via smart connected toys on children and children’s rights. A third contribution that resulted from my research project narrows down on the topic of transparency and how to create a framework for engineers to create products with transparency of automated decision-making in mind. The reason why I decided within these contributions to narrow my research scope was that it became clear to me while writing my first article that in order to contribute meaningfully to the literature I had to either focus on a specific issue (e.g., discrimination, transparency) or on a given context (e.g., children and toy robots). Aside from these publications that result from my DSI research proposal, I seized the unfortunate circumstances we went through to look into the ways the pandemic affected the deployment and creation of (new) technologies or technical applications. While most of the research conducted during this year will be published in 2021, the latter article was recently published in Big Data & Society.

My research has also enabled me to collaborate with DSI fellows as well as with external researchers. First, I had the pleasure to meet Prof. Lampoltshammer at the DSI when meetings in-person were still possible. While we planned a collaboration, unfortunately, due to the pandemic we have so far not been able to collaborate on a research project. I hope that this will change in the future and we will stay in touch. Other fruitful collaborations with DSI fellows or personnel have been made with Moritz Büchi (with a common article on the issue of disillusionment of users about the automated profiling practices on Facebook in development) and Michele Loi (with a common article with the title “Discrimination for the Sake of Fairness”). External collaborations with Simon Mayer (University of St. Gallen), Christoph Lutz and Gemma Newlands (BI Norwegian University), Eduard Fosch-Villaronga and Simone van der Hof (Leiden University), and Gil Scheitlin (UZH) have also enriched my DSI fellowship. Lastly, the fellow exchanges and the possibility to participate in the workshop on Trust in AI have enriched my academic year and I have had the honor to participate at the Digital Days of Switzerland.

Overall, I look back at a very exciting year of research even if the multiple conferences I had hoped to attend in 2020 had to either be cancelled or moved online. Nonetheless, the papers that are currently in the pipeline will be presented in 2021 and the contributions that I had the privilege to work on this year will be available online (open access) soon.

Website: Aurealia's ORCID ID

Prof. Dr. Carmen Tanner

Prof. Dr. Carmen Tanner was DSI Fellow in the fall semester 2019 and was working on the following topic:

Promotion of Personal Integrity via Serious Games and other digital tools

In order to reduce the development and number of questionable business practices in the financial and economic world (such as fraud, deception, corruption), the focus in recent years has been primarily on enhancing regulation and compliance. In contrast, the approach of Moral Intelligence (Tanner & Christen, 2014) assumes that the implementation of corporate values and rules into practice also depends on individual psychological competencies. There is a renewed demand for integrity among managers and employees. However, personal integrity requires a range of perceptual, motivational, decisional and action-related skills. Since a few years, I have been working on the question of how digital tools can be used to promote such moral and psychological skills. In interdisciplinary collaboration and with the DSI we have already developed a first video game (Serious Moral Game), implemented it in education and empirically tested its effectiveness in a research project. The goal during the DSI Fellowship is to consider further developments of this game and/or alternative digital tools to strengthen specifically personal responsibility, decisional ability and courage. A cooperation with a business partner is also planned. 


Center for Responsibility in Finance, Department of Banking and Finance

Chair of Economic Psychology and Leadership Ethics

Prof. Dr. Eva Weber-Guskar

Eva Weber-Guskar was a DSI Fellow from February to July 2019.


The Fellowship focused on research into the ethics of emotional artificial intelligence. AI systems are increasingly becoming direct partners for people in communication and other practices. Examples range from language offerings such as Siri in smartphones, home and shopping assistants such as Alexa, to robots as museum guides or in nursing care for the elderly. This development also includes the fact that the systems are increasingly being equipped with emotional functions (recognising emotions, evoking them in others and simulating one's own expression and reaction behaviour), partly as an aid to communication, partly as an objective in itself. The research question is which problems go hand in hand with such emotionalized AI (EAI), both in a narrow moral philosophical sense and in an ethical-anthropological sense.

In a first lecture on this subject, I examine the various concerns expressed that it might be wrong or problematic as a human being to actually encounter such EAI systems, which suggest feelings, with feelings. I focus on social feelings: compassion and reactive moral emotions. I show that they are epistemically inappropriate and discuss to what extent they can nevertheless be considered desirable for moral reasons. Finally, I refer to how this ambivalence of evaluation can be resolved by considering the possibility of a special epistemic relationship between man and machine: imaginative perception as the basis of emotions. I use the feedback from the various places where I gave the lecture (Zurich, Oldenburg, Bochum) to revise the text for a publication. 


The Fellowship offered me excellent networking conditions. I have participated in the general DSI speed dating, numerous Fellow Exchange Days and Brown Bag Lunches and have become a member of the Digital Ethics Lab of the DSI. Everywhere I have met researchers from whom I could also learn something for my topic. Within the lab I now belong to the group that will organize a workshop on "Trust in Machines". Cooperation is also planned with my next research location, the Weizenbaum Institute for Networked Society.


As a member of the support group, I also advised on the TA Swiss study "Self-driving vehicles in Switzerland", which will be published in autumn 2019. I was invited as a speaker at public evening events at the Kulturverein Rothenburg (Ethische Fragen der Digitalisierung) and at the Literaturhaus Stuttgart (Lange Nacht der KI, "Can you write me a poem, Siri?") and was interviewed by the UZH magazine on the topic " Wir und die Maschinen". On in the blog series DSI-Insights I have summarized my considerations from a research project on the question of how much anonymity in online commentaries is helpful for political discourse. Finally, together with Fellow Dr. Sarah Ebling, I ran a Science Café at Scientifica 2019 on the topic "Can a computer system think? The Turing Test in the Movie and in the Present". 

Website: Prof. Dr. Eva Weber-Guskar

Dr. Lonneke van der Plas

Lonneke van der Plas, senior lecturer in Human Language Technology at the Institute of Linguistics and Language Technology of the University of Malta, was a DSI Fellow in the Fall Semester 2019. Dr. van der Plas is an internationally recognized researcher in the domain of automatic semantic analysis, contributing to the topics such as semantic networks, shallow semantics and multiword expressions. As a DSI Fellow Lonneke van der Plas was working on three main projects.


Coordination of Erasmus+ Strategic Partnerships in the field of higher education. In collaboration with Dr. Tanja Samardzic from the University of Zurich (URPP "Language and Space"), she organised a meeting with potential international project partners from the University of Malta, the University of Belgrade, the University of Bologna, the University of Graz, Clarin ERIC, and the University of Rijeka. The central goal of the project is t tackle skills gaps and mismatches of students studying in language-related disciplines through supporting the development of materials that better meet the learning outcomes needed in the current job market. Technology giants as Google, Amazon, and Facebook all work with language data and the demand for research skills in language-related domains is constantly growing. As our survey shows, university curricula at BA level are rarely oriented towards skills needed methods, and focusing on describing rather than on predicting or explaining linguistic phenomena. As a result, students tend to be poorly prepared for research or industry careers. 

The proposed strategic partnership will introduce an integrated research-oriented perspective into language-related programmes, with a focus on the BA levels. This will enhance students' employability by providing them with the critical skills needed to apply for a wieder range of positions in the labour market, including higher-level ones. 


Creativity and AI. This project is in collaboration with Dr. Michele Loi (Institute of Biomedical Ethics and History of Medicine, Univeristy of Zurich) and focuses on the topic of creativity and AI. This paper aims to put an issue on the agenda of AI ethics that is overlooked in the current discourse. The current discussions are dominated by topics such as trustworthiness and bias, whereas the issue this project focuses on is counter to debate on trustworthiness, in a way. The overuse of currently dominant AI systems that are driven by short-term objectives and optimized for avoiding error leads to a society that loses its diversity and flexibility needed for true progress. The concerns are couched in the discourse around the term anti-fragility and show with some examples what threats current methods used for decision making pose for society. 

Compounding and corpora of non-standard language. Collaborations with Prof. Stark (Co-leader of the URPP "Language and Space", University of Zurich) started at the kick-off meeting, in which the DSI fellow presented her work to the local researchers at the University of Zurich. They are currently collaborating on the topic of liaison in French data. More in particular, they are investigating the correlation between levels of closed-compoundhood as measured in multilingual parallel data and liaison. Liaison is understood as the overt realisation of a latent word-final consonant which (in a specific syntactic/prosodic context) is not pronounced before a following word-initial consonant, but is realised in front of a following word-initial vowel. 

A student of Prof. Stark will annotate French audio data on which a BSc student from Malta will apply statistical processing in order to test several hypotheses regarding liaison and the lexical association between words in the phrase, as formulated in Prof. Stark's earlier work. 

Website: Dr. Lonneke van der Plas

Dr. Bruno Wüest

Website: Dr. Bruno Wüest

Dr. Kristina Yordanova

Dr. Kristina Yordanova was a guest at the DSI and the UFSP Dynamics of Healthy Ageing in July and August 2018.

The main project I pursued within my DSI fellowship was the development of automated methods for situation knowledge extraction and interpretation of human behaviour. This involves the investigation of both different types of information sources and different computational solutions for the interpretation of heterogeneous data. In that context, I was able to propose a workflow for the automatic identification of social behaviours and environments from transcripts of daily conversations. The initial results suggested that it is possible to apply automated methods for coding transcripts of daily conversations with relevant psychological variables.  Another aspect of the project was identifying relevant factors that describe healthy ageing from a situation-aware perspective. These factors can later be used to reason about a person’s daily life, the evolvement of their situation and the effect these factors have on the long-term behaviour associated with healthy ageing.

Website: Dr. Kristina Yordanova