Voici les éléments 1 - 9 sur 9
  • Publication
    Accès libre
    Generative AI-Enabled Conversational Interaction to Support Self-Directed Learning Experiences in Transversal Computational Thinking
    (2024-07) ; ;
    Juan Carlos Farah
    ;
    Denis Gillet
    ;
    As computational thinking (CT) becomes increasingly acknowledged as an important skill in education, self-directed learning (SDL) emerges as a key strategy for developing this capability. The advent of generative AI (GenAI) conversational agents has disrupted the landscape of SDL. However, many questions still arise about several user experience aspects of these agents. This paper focuses on two of these questions: personalization and long-term support. As such, the first part of this study explores the effectiveness of personalizing GenAI through prompt-tuning using a CT-based prompt for solving programming challenges. The second part focuses on identifying the strengths and weaknesses of a GenAI model in a semester-long programming project. Our findings indicate that while prompt-tuning could hinder ease of use and perceived learning assistance, it might lead to higher learning outcomes. Results from a thematic analysis also indicate that GenAI is useful for programming and debugging, but it presents challenges such as over-reliance and diminishing utility over time.
  • Publication
    Accès libre
    Designing digital choice architecture : studies on the behavioral and attitudinal impact of digital nudges
    (Université de Neuchâtel, 2024-02-22) ; ;
    Cette thèse explore l'espace de conception des nudges numériques et leurs effets sur les utilisateurs dans trois domaines d'application : la vie privée en ligne, l'utilisation des médias sociaux et les dons de charité en ligne. Après le chapitre d'introduction, une revue systématique de la littérature liée aux études empiriques sur les nudges numériques est menée, révélant le paysage actuel de la recherche. Ensuite, une expérience en ligne et une étude descriptive sur le terrain sont utilisées pour démontrer l'efficacité des nudges numériques dans la sensibilisation à la protection de la vie privée en ligne pour les utilisateurs de certaines plateformes digitales. Le chapitre suivant utilise une expérience de terrain, menée en ligne, pour modifier l'architecture des choix numériques d'une plateforme de médias sociaux populaire. Cette étude évalue les régimes restreints des fils d'actualité et leur impact sur les comportements et les expériences des utilisateurs. Le cinquième chapitre est une expérience factorielle en ligne concernant les dons de charité avec des enjeux monétaires réels. Cette dernière étude examine systématiquement les combinaisons de plusieurs nudges numériques sur les comportements de don en ligne des utilisateurs et leurs attitudes associées en matière de manipulation et d'autonomie. La thèse présente plusieurs contributions liées aux nudges numériques et à l'architecture de choix. L'examen systématique met en lumière plusieurs pistes de recherche ouvertes pour le domaine qui appellent à des recherches futures. L'expérimentation liée à la confidentialité en ligne montre que les nudges numériques pourraient améliorer les interactions des utilisateurs avec les conditions et les politiques de confidentialité, sans toutefois renforcer leur mémorisation de ce qu'ils ont accepté. L'expérience sur le terrain concernant les régimes des fils d'actualité des médias sociaux démontre leur impact sur le temps passé sur la plateforme de médias sociaux, mais ces interventions de conception s'accompagnent d'expériences positives et négatives pour l'utilisateur. L'expérience factorielle en ligne relative aux dons de charité révèle que certaines combinaisons de nudges numériques augmentent les dons, mais ces nudges suscitent également des inquiétudes chez les utilisateurs quant aux menaces pesant sur leur autonomie et leur sentiment de manipulation. Ensemble, les chapitres de la thèse mettent en évidence l'impact complexe des nudges numériques sur le comportement et les attitudes dans plusieurs domaines d'application pertinents. ABSTRACT: This dissertation explores the design space of digital nudges and their effects on users in three application domains: online privacy, social media usage, and online charitable giving. After the introductory chapter, a systematic review of empirical studies on digital nudging is conducted, uncovering the current research landscape. Next, an online experiment and a descriptive field study is used to demonstrate the effectiveness of digital nudges for enhancing the online privacy awareness of users for popular online services. The next chapter uses an online field experiment to alter the digital choice architecture of a popular social media platform. This study evaluates restricted newsfeed diets and their impact on user behaviors and experiences. The fifth chapter is an online factorial experiment concerning charitable giving with real monetary stakes. This last study systematically investigates combinations of several digital nudges on users' online donation behaviors and their associated attitudes around manipulation and autonomy. The dissertation extends several contributions related to digital nudging and choice architecture. The systematic review illuminates several open research avenues for the domain at large that call for future research. The experiment related to online privacy awareness shows that digital nudges could elevate user interactions with terms and privacy policies, although without bolstering users' recollections of what they have agreed to. The field experiment concerning social media newsfeed diets demonstrates their impact regarding time spent on the social media platform, but these design interventions come with both positive and negative user experiences. The online factorial experiment related to charitable giving reveals that some combinations of digital nudges increase donations, but these particular nudges also invoke users' concerns about autonomy threats and an experienced sense of manipulation. Together, the chapters of the dissertation highlight the complex behavioral and attitudinal impact of digital nudges across several relevant application domains.
  • Publication
    Accès libre
    On the Impact of Digital Boosts on Perceived Stress in a Self-Regulated Learning Experiment
    Self-regulated learning (SRL) has been adopted as a successful strategy for promoting deeper learning and improving academic performance. In this context, digital boosts have been used to empower learners by expanding their competencies and helping them reach their objectives. However, existing literature has primarily focused on the academic performance-related outcomes of digital boosts, while their potential effects on emotional and psychological aspects like stress and well-being remain comparatively under-explored. In this study, we address this gap by studying the impact of digital boosts on perceived stress, in addition to study time. We have designed a digital SRL support system, through which we have delivered digital feedback boosts. To evaluate this system, we conducted a pilot study with 60 university students. Our digital boosts have helped students keep a steady study time. However, they have caused an increase in perceived stress, especially among students who did not attain their study time plans.
  • Publication
    Accès libre
    Leveraging ChatGPT to Enhance Computational Thinking Learning Experiences
    Given the pervasive reliance on technology in modern society, teaching Computational Thinking (CT) abilities is becoming increasingly relevant. These abilities, such as modeling and coding, have become crucial for a larger audience of students, not only those who wish to become software engineers or computer scientists. Recent advances in Large Language Models (LLMs), such as ChatGPT, provide powerful assistance to complete computational tasks, by simplifying code generation and debugging, and potentially enhancing interactive learning. However, it is not clear if these advances make CT tasks more accessible and inclusive for all students, or if they further contribute to a digital skills divide, favoring the top students. To address this gap, we have created and evaluated a novel learning scenario for transversal CT skills that leveraged LLMs as assistants. We conducted an exploratory field study during the spring semester of 2022, to assess the effectiveness and user experience of LLM-augmented learning. Our results indicate that the usage of ChatGPT as a learning assistant improves learning outcomes. Furthermore, contrary to our predictions, the usage of ChatGPT by students does not depend on prior CT capabilities and as such does not seem to exacerbate prior inequalities.
  • Publication
    Accès libre
    Digital Nudges for Privacy Awareness: From consent to informed consent?
    (AIS, 2020-6-15) ;
    Gjerlufsen, Tony
    ;
    Maingot, Paul
    ;
    ;
    Maintaining a private life in our digital world is gradually becoming harder. With Internet services having ever increasing access to personal data, it is crucial to raise user awareness about what privacy guarantees they offer. Regulations have recently been enacted such as the European General Data Privacy Regulation (GDPR). Yet, online service providers still have terms and privacy policies to which users tend to agree without ever viewing or reading them. By using digital nudges, this paper explores how small changes in the choice architecture can be designed to increase the informed consent and privacy awareness of users. The results from a double-blind online experiment (n = 183) show that phrasing the agreement differently and providing a highlights alternative to the existing quick-join choice architecture can significantly increase the number of users who view and read the terms and privacy policy. However, these digital nudges seem to not increase the users’ recollection of what they have agreed to. The experimental results are complemented by a field test using one of the proposed designs in the IKEA Place app (n = 81’431).