In his article “Why AI Won’t Democratize Education”, Wieczorek (2025) challenges the techno-optimistic narrative surrounding intelligent tutoring systems. He argues that these systems, despite their promise of personalized feedback and educational efficiency, will not democratize education from a pragmatist perspective. His central critique focuses on a fundamental limitation: intelligent tutoring systems do not cultivate the capacity for human communication and collaboration. Skills essential to democratic participation and citizenship. This commentary extends Wieczorek’s argument by adding that unconstrained large-language models (chatbots) undermine education in general, not merely democratic education. Drawing on the concept of “resistance” articulated by Philippe Meirieu, combined with arguments from cognitive psychology, we argue why the user-friendliness of unconstrained chatbots is pedagogically problematic. We explore whether chatbots and pedagogy can be commensurable by discussing constrained AI applications that reintroduce productive friction while enhancing rather than replacing human judgment and teaching. Ultimately, we argue, there might be a place for constrained AI in education, but the current evidence does not warrant it.
Studies into affective publics often involve textual communication. However, emotive communication is increasingly visual. This study zooms in on the representation of the suffering other in seven re-workings of the Alan Kurdi photographs that resonated significantly on Instagram. Chouliaraki’s concept of post-humanitarian solidarity in The Ironic Spectator (2013) is used as a theoretical framework to analyse the content of re-worked images and their post captions. Her concept outlines how distant sufferers tend to be rendered invisible due to the self-reflexive nature of contemporary solidarity. This self-reflexivity gets in the way of solidarity for others unlike us. The study found that, although the sufferer is visually present in almost all re-worked images, the suffering is ‘replaced’ by emotions or political views of the creators. Both Chouliaraki’s ‘distant other’ as well as Markham’s similar other are ways to visually (re)construct the tragedy of Alan Kurdi and the refugee crisis in general. This study adds to this an understanding of how Instagram users, while visually constructing a similar or distant other, also write themselves – often their personal feelings – into such images. Their public, other Instagram users, engages in self-reflexivity by liking such re-workings, aligning with the communicated emotions or political views conveyed. In this way, the platform ‘like feature’ intensifies the self-reflexive nature of contemporary solidarity.
As artificial intelligence (AI) systems are increasingly deployed in high-risk financial decision-making contexts, the demand for transparency and interpretability becomes critical. Explainable AI (XAI) has emerged as a key research domain addressing these needs. While most existing XAI studies emphasize objective quality measures such as correctness and completeness of explanations, they often overlook the role of end-user requirements and the broader ecosystem of stakeholders. This study presents a human-centered evaluation of different visual explanation designs in financial AI applications, assessing their effectiveness. A two-phase mixed-method evaluation was conducted, combining user studies with end-users and a stakeholder workshop, to rank visual prototypes across four explanation types: feature importance, counterfactuals, contrastive/similar examples, and rule-based explanations. A key finding is the divergence between end-users and other stakeholders—including compliance officers, XAI consultants, and developers—with end-users indicating a preference for concise, contextually visual explanations (e.g., small sets of decision rules or risk plots relative to similar cases), while other stakeholders often favor more complete, technically detailed representations. This highlights a critical trade-off between interpretability and completeness. This suggests that visual encoding choices may affect the effectiveness of AI explanations across different stakeholder groups.
Malnutrition is a multifactorial and complex condition with significant consequences for recovery, functional outcomes, and healthcare systems. Research in malnutrition is often limited by single-component interventions, heterogeneous study designs, and variable outcome measures. This perspective paper introduces a practical guiding framework for clinical nutrition research, emphasizing interdisciplinary, multifactorial approaches, co-designed interventions, and pragmatic, adaptive study designs. Evidence from several trials demonstrates that individualized nutritional support delivered by multidisciplinary teams improves clinical outcomes, yet challenges remain in recruitment, adherence, and balancing intervention intensity with patient burden. The framework provides a structured approach to intervention development, outcome selection, and implementation, while remaining flexible to accommodate innovation, context-specific adaptation, and emerging outcome measures. By integrating lessons from prior trials, globally, and promoting systematic reporting and feasibility assessment, this framework aims to enhance the design, comparability, and translational impact of future research in clinical nutrition in older and other clinically vulnerable populations. Adoption of such a framework can guide research prioritization, optimize intervention delivery, and ultimately improve patient recovery and quality of life.
Artificial Intelligence (AI) agents personalize their responses by tailoring explanations to users' backgrounds, interests, and prior interactions, referred to as contextualization. Personalization has been identified as a persuasive strategy in politics or in marketing. However, the persuasive effect of contextualization in everyday tasks, where users often lack prior knowledge, remains unclear. We conducted a 2×2 between-subjects experiment (N = 380) examining how contextualization, combined with conversational warmth, shapes reliance and persuasiveness of an AI assistant arguing against expert recommendations. Our findings reveal that contextualization reduces the persuasive power of AI, but its combination with warmth restores persuasiveness through a crossover interaction. Reliance on AI is present across conditions and is invariant to the conversational design. Trust strongly predicts both persuasion and reliance, yet neither contextualization nor warmth operates through trust. AI literacy decouples trust from behavior: more literate users report lower trust in the assistant, yet are more persuaded and more reliant on its advice. These results suggest that users are prone to deferring to AI agents over human expert judgment; however, interface-level conversational design choices have a limited role in shaping the behavior.
