University of St. Gallen

The widespread emergence of artificial intelligence (AI) is expected to support the legal sector in its daily work and, most importantly, relieve it of laborious tasks. Both the promise and the temptation are very significant. Therefore, it is essential to adopt a critical perspective on this matter. This is not due to being against technology but rather to the fact that its use requires effective risk management that surpasses mere reliance on technology, or rather, a zero tolerance for errors. It is crucial to strategically position AI in the value chain of legal service providers and reflect it in the chosen business model. In addition, the impact on leadership must be analysed because, at the end of the day, there are still human beings at work. And it is imperative to focus on potential convenience traps: the loss of human creativity and negotiation skills, a reduction in valuable training exercises, a lack of plausibility checks for AI solutions, and a deterioration in service quality. Lawyers will have to face the fact that they have to think about the use of AI if they want to remain relevant to their internal and external customers. Legal professionals are strongly encouraged to pursue a well-informed and responsible approach towards AI with the aim of safeguarding their professional integrity and prepare themselves for potential changes at an early stage.

Open-texture—e.g. vague, ambiguous, under-specified, or abstract terms—in regulatory documents lead to inconsistent interpretation, and are an obstacle to the automatic processing of regulation by computers. Identifying which parts of a legal text fall under open-texture is therefore a necessary requirement to make progress in automating the law. In this paper, we propose that large language models (LLMs) might provide an effective way to automatically detect open-texture in legal texts. We first investigate the obstacles by situating open-texture in the broader literature, and we test the hypothesis using two different LLMs—the proprietary gpt-3.5-turbo and the open-source llama-2-70b-chat—for the task of identifying open-texture in the General Data Protection Regulation. We evaluate their performance by asking 12 annotators to assess their output. We find, overall, that gpt-3.5-turbo overperforms llama-2-70b-chat on F1-scores (0.84 vs 0.67), and its high F1-score could make it a suitable alternative, or complement, to using human annotators. We also test the sensitivity of the findings against four further LLMs combined with six different prompts, and replicate a finding that there is low agreement between annotators when it comes to the identification of open-texture. We conclude the article by discussing the subjectivity of open-texture, the lessons to draw when testing for open-texture, and the consequences of using LLMs in the legal domain.

Objective Global spending on diabetes care soared to $966 billion in 2021, a 316% surge over the past 15 years. This sharp increase underscores a need for more efficient and cost-effective care strategies. Value-based care (VBC), which prioritizes patient outcomes while controlling expenses, presents a promising solution. However, its real-world implementation remains challenging, particularly in diabetes care. This study examines SwissDiabeter, a proposed diabetes clinic initiative in Switzerland inspired by a Dutch VBC-based Diabeter clinic. We examine key barriers and facilitators during Diabeter's implementation in the Netherlands and assess forthcoming challenges and enablers for SwissDiabeter in Switzerland. Methods We employ a deep, extensive embedded single-case design conducting 27 interviews with healthcare professionals, insurers, and patient groups in Switzerland and the Netherlands. The main interview data were complemented by various secondary sources to enhance contextual comprehension, widen perspectives, and validate findings. Results We identify four key factors for successful VBC adoption: leadership in driving change, financial restructuring, operational improvements, and enabling digital technologies. We next derive practical recommendations to guide the implementation of value-based diabetes care, redesigning financial incentives for healthcare providers, partnering up with key stakeholders such as insurers or policy makers, and measuring outcomes on a voluntary and anonymous basis. Conclusion This study enhances the global discourse on VBC by analyzing key barriers and facilitators in implementing SwissDiabeter, drawing insights from the Diabeter model in the Netherlands. Our findings highlight the need for strong leadership, financial incentives, digital infrastructure, and interdisciplinary collaboration to drive outcome-driven care. Beyond diabetes, these insights provide a framework for scaling VBC across chronic disease management, promoting cost-effective, high-quality healthcare.

Background and objective Dementia prevention has been recognized as a top priority by public health authorities due to the lack of a reversible cure. In this regard, digital dementia-preventive lifestyle services (DDLS) emerge as potentially pivotal services, aiming to address modifiable risk factors on a large scale. This study aims to identify the top-funded companies offering DDLS and evaluate their clinical evidence to gain insights into the international service landscape. Methods A systematic screening of two financial databases (Pitchbook and Crunchbase) was conducted. Corresponding published clinical evidence was collected through a systematic literature review and analyzed regarding study purpose, results, quality of results, and level of clinical evidence. Findings The ten top-funded companies offering DDLS received a total funding of EUR 128.52 million, of which three companies collected more than 75%. Clinical evidence was limited due to only nine eligible publications, small clinical subject groups, the absence of longitudinal study designs, and no direct evidence of dementia prevention. Conclusion Our study shows that the level of funding received by companies does not reflect the clinical effectiveness of DDLS. The study serves as an initial step toward understanding how DDLS are currently evaluated in today’s market and highlights the need for a more rigorous evaluation of DDLS effectiveness.

Providing argumentation feedback is considered helpful for students preparing to work in collaborative environments, helping them with writing higher-quality argumentative texts. Domain-independent natural language processing (NLP) methods, such as generative models, can utilize learner errors and fallacies in argumentation learning to help students write better argumentative texts. To test this, we collect design requirements, and then design and implement two different versions of our system called ALure to improve the students’ argumentation skills. We test how ALure helps students learn argumentation in a university lecture with 305 students and compare the learning gains of the two versions of ALure with a control group using video tutoring. We find and discuss the differences of learning gains in argument structure and fallacies in both groups after using ALure, as well as the control group. Our results shed light on the applicability of computer-supported systems using recent advances in NLP to help students in learning argumentation as a necessary skill for collaborative working settings.

Plain English Summary Action-oriented entrepreneurship training can reduce unemployment among young adults in sub-Saharan Africa by fostering a stabilized action-motivation pattern that forms a path to self-employment. The study presents a path-centric account that describes how entrepreneurial motivation and action influence and stabilize each other after action-oriented entrepreneurship training. We ran field experiments with more than 1000 students enrolled at universities in sub-Saharan Africa. Participation in action-oriented entrepreneurship training increased the likelihood of being self-employed 1 year after training by 30%. This could be explained by a reciprocal and mutually stabilizing relationship between entrepreneurial motivation and action after training. Thus, action-oriented entrepreneurship training should be considered an effective means to reduce unemployment among young adults in sub-Saharan Africa, because it promotes an action-motivation path leading to self-employment.

Background Europe’s healthcare systems face a triple burden: the rise of non-communicable diseases (NCDs), an aging population, and a shortage of healthcare professionals. NCDs, the leading causes of death, disproportionately affect older adults, placing significant pressure on healthcare services. By 2050, nearly 30% of Europe’s population will be aged 65 or older, up from 20% in 2023. These challenges demand urgent solutions to sustain healthcare systems. Patient-facing digital health technologies (DHTs), such as Digital Diagnostics and Digital Therapeutics, offer promising tools to address this burden by empowering patient self-management, reducing strain on healthcare professionals, and enhancing system efficiency. Despite their potential, the scaling and adoption of DHTs remain limited. This study investigates: (RQ1) What key factors drive success across different patient-facing DHT categories? and (RQ2) How can companies implement these factors? Methods Following COREQ guidelines, we conducted semi-structured interviews with 29 executives and founders of European DHT companies targeting NCDs. Participants were identified using PitchBook, focusing on revenue-generating companies with over 20 employees. Virtual interviews were conducted in English between May and September 2024, lasting an average of 28 min (range: 21–40). Data saturation determined the sample size. Thematic analysis was performed, with two researchers independently coding the data to ensure reliability. Success factors were categorized as internal (e.g., employees) or external (e.g., partnerships). Ethical approval was obtained, and data was anonymized. A follow-up survey (n = 27) was conducted to confirm our findings. Results We identified 18 success factors for scaling patient-facing DHTs. Health & Wellness companies prioritized business model flexibility, while Digital Therapeutics relied on regulatory compliance. Validation of health impact was critical across categories, emphasized by all respondents in Digital Diagnostics and Digital Therapeutics. Other key factors included customer awareness, strategic partnerships, and investor alignment, highlighting the importance of tailored growth strategies. Conclusion This study provides structured guidance for scaling patient-facing DHTs, emphasizing category-specific strategies aligned with operational, regulatory, and consumer demands. It offers actionable recommendations for founders and executives to allocate resources effectively and adapt to diverse market contexts. By addressing the unique challenges of scaling DHTs, this work contributes to advancing digital health research and improving healthcare system resilience.

Background Travel-related psychiatric disorders range from anxiety disorders to mood disorders, substance abuse, and psychosis. Various travel-associated factors such as dehydration, time shifts, changes in social structures or stress factors are discussed for these disorders. There is a lack of knowledge concerning the quality and outcome of psychiatric treatment in travelers hospitalized abroad. This study is the first to compare outcome of treatment in psychiatric travelers to domestic patients. Methods We analyzed electronic health records of travelers in the Psychiatric University Hospital Zurich from January 2013 to December 2020. Each traveler was matched with one Swiss national and one migrant using propensity score matching. Results Travelers showed inferior CGI-I scores at discharge (F(2,969) = 5.72; p = 0.003). The length of stay was shorter (F(2,969) = 38.74:p < 0.001) for travelers (9.69 ± 14.31) than for Swiss nationals (24.69 ± 29.42) and migrants (24.74 ± 28.62). The transfer rate to another hospital was higher (X²(2,972) = 50.85: p < 0.001) for travelers (79, 29.4%) than for Swiss nationals (25, 7.7%) or migrants (26, 8.0%). Conclusions Psychiatric treatments of hospitalized travelers showed a lower symptom improvement while presenting a more severe overall condition at discharge. Length of stay was shorter compared to domestic patients. Admission of travelers was initiated involuntarily more frequently. This most closely reflects the theory that travelers are typically hospitalized in severe emergencies and are promptly discharged or repatriated after an initial treatment response has been achieved.

Myriads of sustainability measurement schemes are in place to assess the societal impact of firms and other organizations. While these schemes aim to encourage positive social and environmental outcomes, they often fall short. We identify interest comprehensiveness, cognitive accuracy, and practical feasibility as the aspired dimensions of these schemes to argue why they are bound to be imperfect. Next to difficulties within each dimension, sustainability measurement schemes face inherent tensions between the three dimensions, which we conceptualize as commensuration, scope, and scale ambivalence. We introduce the three papers for this themed section that develop salient frameworks and rich empirical insights into these tensions as well as the upsides of sustainability measurement schemes from very different theoretical and methodical perspectives. We conclude that while the inherent tensions within and between the aspired dimensions render the accomplishment of perfect schemes a mission impossible, there are viable solutions that pave the way to more effective sustainability measurement schemes. Promising pathways to which organizational scholars can make significant contributions include improving the governance of schemes to enhance transparency and balance power, using innovative data to more accurately and exhaustively feed these schemes, streamlining the numerous schemes to remove redundancy, and simplifying schemes to make them more robust.

Mobile navigation applications are good at providing efficient navigation instructions. However, they currently lack the capability to facilitate free exploration. Therefore, users are limited to encountering only places close to the shortest paths, neglecting places that could diversify navigation and foster spatial learning. To better understand what characteristics places have that users like to explore we collected a dataset with a mobile application that encourages free exploration using gamification (n = 39, t = 455 days, 106.50 𝑘𝑚2). Using OpenStreetMap data, we found highly frequented freely explored places comprising office, educational, retail, touristic and commercial places. When comparing the characteristics of the freely explored places to those along the shortest path, those categories were different. Based on our findings, we propose that implementing more diverse routing algorithms can enhance navigation diversity, improve spatial learning, and optimise the utilisation of urban spaces for travel.

In this study, we explore the potential of AI-generated podcasts as an educational tool in the evolving landscape of learning media. Podcasts have grown increasingly relevant in education due to their accessibility and ability to integrate learning into everyday life. With the advent of generative artificial intelligence (AI), there is a unique opportunity for scalable and adaptable creation of learning media. However, with novel technology, there also come new challenges. Thus, we developed fine-tuned AI-generated podcasts using Google NotebookLM, our course materials, and a custom prompt. We conducted a one-month explorative evaluation in the field using a qualitative diary study. Our study reveals that students find the podcasts beneficial for flexible everyday learning but also point toward challenges like a lack of emotional engagement and Permission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citation on the first page. Copyrights for third-party components of this work must be honored. For all other uses, contact the owner/author(s). technical non-English language issues. In sum, our study highlights the current benefits and challenges of AI-generated podcasts and presents an agenda for future research.

While the formation of higher education institution (HEI) alliances is usually driven by purposes of internationalization, the European Union has recently turned to HEI alliances as problem solving tools for societal challenges. The formation of HEI alliances and their missions result from the complex interplay of multiple levels, actors, and issues and notably depend on the political agenda of their source of funding. Since the combat against climate change is assuming a central role in political agendas at (sub)national and supranational level, it is expected to also shape the logics of HEI alliance formation. This contribution asks how HEI alliances are affected by the impetus for a stronger contribution by HEIs to tackle climate change. The analysis investigates the role of climate change for the missions and cooperation results of HEI alliances in the European Education Area. In terms of missions, the chapter finds that HEI alliances cooperate in climate issues due to the profiles of the involved HEIs and due to societal demands and funding stipulations. Regarding cooperation results, the analysis shows that HEI alliances primarily incorporate climate change action into short-term joint teaching formats and into the exchange of researchers with shared interests. Moreover, HEI alliances serve to exchange practices to make campuses and different aspects of university life such as student mobility more sustainable.