Accumulation of legal design knowledge in software development

Legal knowledge is becoming an increasingly important part of software development. New stricter data protection regulations such as the GDPR increase the pressure on developers to implement legal requirements. Design patterns are an established tool in software development to capture proven solutions to recurring problems. Thus, design patterns can also be a way to share legal knowledge with developers.

The research project investigates how developers of smart personal assistants such as Amazon Alexa can be supported in developing the systems as legally as possible and in accordance with the latest data protection regulations.

Possible theses can deal with:

• GDPR compliant smart personal assistants
• Design patterns to codify legal design knowledge
• Development of frameworks to support developers in solving legal challenges
• Qualitative and quantitative studies with developers
• Research on how developers can best improve their skills and extend their legal knowledge

Start immediately or by arrangement.

When using digital technologies and making decisions on digital environments, e.g., booking a flight, various data traces are left behind for collection, storage and analysis. Innovative solutions for information systems are needed that mitigate privacy risks and foster information privacy. One mechanism to achieve this is the use of privacy nudges. Nudges are a concept from behavioral economics to influence individual’s decision making. Nudges intend to alter people’s behavior in a predictable way without forbidding any options or significantly changing their economic incentives.

As part of a greater research project, this study seeks to create and evaluate a process model for designing digital nudges that foster privacy sensitive behavior.

Start immediately or by arrangement

Why:
The generation “Millennials” request to receive more feedback from their organizations to develop and learn. This is reflected by a growing number of digital feedback apps which facilitate high-frequency feedback exchange. However, the effect of feedback has hardly been studied on an organizational level due to complexity. Therefore, the goal of this master thesis is to analyze organizational feedback exchange with an agent-based simulation model. This simulation method is well known from Covid-19 in which researchers try to predict events and behaviors of people under certain conditions.

How:
You don’t need to start from scratch, you can extend and build our existing model (which won the most innovative paper award)

We use the software NetLogo, which has an easy to understand syntax (no programming skills required)
Key-Facts:
The collaboration will take place in close supervision, with regular review meetings, feedback sessions, etc.
You will work in an international research team (Temple Uni-USA and HSG)
The thesis can be started immediately, but should be completed within the next ±6 months

Literature on Agent-Based Modeling:

Davis, Jason P.; Eisenhardt, Kathleen M.; Bingham, Christopher B. (2007): Developing Theory Through Simulation Methods.

Beese, Jannis; Haki, M. Kazem; Aier, Stephan; Winter, Robert (2019): Simulation-based research in information systems.

Our previous work

AI-based information systems offer the potential to augment human decision-making across a variety of use cases, such as credit scoring, local policing or cancer detection. However, such systems offer outcomes that are not fully predictable nor explainable, and bring along a certain probability of inaccuracy. Current research is incongruent on whether and under which circumstances users actually accept and trust such erroneous recommendations. More so, the existing literature body has mostly focused on the positive outcomes of AI-based information systems. Which individual and technological factors promote or prevent the acceptance of erroneous recommendations?

The goal of this thesis is to analyse social and psychological theory (i.e., decision-making) relevant to the area of research described above. In a potential subsequent step, the student is asked  to deduct relevant impediments of recommendation acceptance from a socio-technical design perspective.

Start immediately or by arrangement

From smart home devices such as Google Home or Amazon Alexa, voice control systems in vehicles and half of all search requests being performed by voice  – technology has shifted towards voice-based interfaces within the past years. These hands- and eyes-free interfaces have the potential to optimize work processes (think medical surgery) or reach new customer groups (think visually impaired, elderly). But what happens to those target groups who cannot be understood? For instance, accuracy of search results can differ depending on whether an algorithm recognizes your accent.

As part of a greater research project, this study seeks to identify the unintended implications the wide-spread deployment of commercial voice interfaces bring along. The student might investigate how language variants or accents (i.e., Hochdeutsch vs Schweizerdeutsch) influence the interaction with a voice-based interface as part of an empirical study.

Start immediately or by arrangement

In der Unternehmenspraxis findet sich inzwischen ein grosser Markt für Tests, welche die Agilität der Mitarbeitenden erfassen sollen. In dieser Masterarbeit sollen systematisch die verwendeten Testverfahren überblickt und deren Güte auf Basis von methodischem Knowhow überprüft werden. Auf diesen Erkenntnissen sollen Handlungsempfehlungen für Vorgehensweisen bei der Vermessung der Agilität von Mitarbeitenden abgegeben werden.

Voraussetzung: gründliches methodische Knowhow

Start per sofort oder nach Vereinbarung

Different topics possible at any time such as NLP, ML, Process Mining, Design, Chatbot, Voice-Interfaces – more information through ai-for-education.com

In these master theses, the students should build upon recent advantages in the field of Machine Learning and Natural Language Processing to investigate the potential of, conceptualize or design new approaches in the fields of human-computer interaction and artificial intelligence for education. The aim is to design new forms of interaction and experience to revolutionize the individual learning experiences of students (e.g., through chatbots or adaptive learning feedback). The student can rely on a set of already existing data, code, tools, and knowledge in order to work with the Ph.D. student together on a novel learning tool for a real-world evaluation. I offer close supervision, individual coaching, and the opportunity to not only develop practical, as well as theoretical skills in the area of Data Science, Machine Learning, and Natural Language Processing but also to publish a scientific paper together at the end of the thesis.

Start: by arrangement

Bei der Umsetzung von KI-Projekten spielt Bias eine grosse Rolle.

Unter Bias versteht man eine systematische Abweichung der Ergebnisse eines Algorithmus von den gewünschten Ergebnissen. Dadurch können KI-Projekte langfristig Schaden nehmen und Vertrauen in die KI-Lösung verspielt werden. Verschiedene Bias können in jeder Phase des Projektes auftreten, weshalb es wichtig ist, sich in jeder Phase des Projektes der möglichen Bias bewusst zu sein. Ist ein Bias identifiziert, kann er mit speziellen, auf den jeweiligen Bias angepassten Mitigation Strategys behandelt werden.

Für die Entwicklung eines Risk Management Tools, das Projektleitern hilft, strukturiert gegen Bias vorzugehen, suchen wir einen Masteranden/ eine Masterandin.