ZoekresultatenZoekresultaten
Volg ons op
Volg ons op
Types
0Instelling
26Publicatiejaar
6Producttype
19Bestandstype
9Taal
5Onderzoeksthema's
13Publicaties met bestand/URL
2Projectstatus
3This article investigates gender bias in narratives generated by Large Language Models (LLMs) through a two-phase study. Building on our existing work in narrative generation, we employ a structured methodology to analyze the influence of protagonist gender on both the generation and classification of fictional stories. In Phase 1, factual narratives were generated using six LLMs, guided by predefined narrative structures (Hero's Journey and Heroine's Journey). Gender bias was quantified through specialized metrics and statistical analyses, revealing significant disparities in protagonist gender distribution and associations with narrative archetypes. In Phase 2, counterfactual narratives were constructed by altering the protagonists’ genders while preserving all other narrative elements. These narratives were then classified by the same LLMs to assess how gender influences their interpretation of narrative structures. Results indicate that LLMs exhibit difficulty in disentangling the protagonist's gender from the narrative structure, often using gender as a heuristic to classify stories. Male protagonists in emotionally driven narratives were frequently misclassified as following the Heroine's Journey, while female protagonists in logic-driven conflicts were misclassified as adhering to the Hero's Journey. These findings provide empirical evidence of embedded gender biases in LLM-generated narratives, highlighting the need for bias mitigation strategies in AI-driven storytelling to promote diversity and inclusivity in computational narrative generation.
ARTIKEL
Het woord ‘bias’ komt naar voren in zowel maatschappelijk als wetenschappelijk debat over de inzet van artificiële intelligentie (ai). Het verwijst doorgaans naar een vooroordeel dat iets of iemand vaak onbedoeld heeft. Wanneer dit vooroordeel leidt tot een afwijking in besluitvorming vergeleken met een situatie wanneer dit vooroordeel er niet zou zijn, dan is een bias doorgaans onwenselijk.
ANDERSOORTIG MATERIAAL
It is crucial that ASR systems can handle the wide range of variations in speech of speakers from different demographic groups, with different speaking styles, and of speakers with (dis)abilities. A potential quality-of-service harm arises when ASR systems do not perform equally well for everyone. ASR systems may exhibit bias against certain types of speech, such as non-native accents, different age groups and gender. In this study, we evaluate two widely-used neural network-based architectures: Wav2vec2 and Whisper on potential biases for Dutch speakers. We used the Dutch speech corpus JASMIN as a test set containing read and conversational speech in a human-machine interaction setting. The results reveal a significant bias against non-natives, children and elderly and some regional dialects. The ASR systems generally perform slightly better for women than for men.
PAPER
Narrative structures such as the Hero’s Journey and Heroine’s Journey have long influenced how characters, themes, and roles are portrayed in storytelling. When used to guide narrative generation in systems powered by Large Language Models (LLMs), these structures may interact with model-internal biases, reinforcing traditional gender norms. This workshop examines how protagonist gender and narrative structure shape storytelling outcomes in LLM-based storytelling systems. Through hands-on experiments and guided analysis, participants will explore gender representation in LLM-generated stories, perform counterfactual modifications, and evaluate how narrative interpretations shift when character gender is altered. The workshop aims to foster interdisciplinary collaborations, inspire novel methodologies, and advance research on fair and inclusive AI-driven storytelling in games and interactive media.
CONFERENTIEBIJDRAGE
Questions we care about (objectives)
This study contributes to the body of knowledge of entrepreneurial educators’ strategies to work towards gender inclusion in entrepreneurial education (EE). By illustrating how gender stereotypes and gender bias are constructed and reconstructed in EE. An important insight is taken from the study by (Jones, 2014), this author shows that entrepreneurs are presented in EE as a homogeneous group with similar character traits and an equal 'entrepreneurial mindset' (p. 244). This description portraits a right way to be an entrepreneur which is traditionally been associated with a white western man, masculine behaviour and masculine abilities (Jones, 2014). As a consequence, a paradox in EE appears; training and education reduces the gender gap on the one hand (Cheraghi and Schøtt, 2015), but at the same time it is also the place where the gender gap is maintained because these gender stereotypes are intertwined in this training and education (Korhonen, Komulainen and Räty, 2012). The aim of this paper is to use Social Role Theory in order to better understand the dynamics of gender in EE. This leads to the following research questions: what are the main mechanisms that contribute to gender-inclusive entrepreneurship education (EE) and how are the different mechanisms that contribute to gender inclusive entrepreneurship education (EE) integrated into the current curriculum?
Approach
In-depth interviews with 12 lecturers that teach/coach in the field of EE from across various faculties in a large University for Applied Sciences combined with newsletters created by one of largest EE programs at this institution (267 pages in 2020) is carried out. A semi-structured questionnaire was used to guide the in-depth interviews. Subsequently, discourse analysis gives insight into gendering in EE at the investigated institution.
Results
The main findings suggests that when questioned about the topic all respondents stressed the importance of gender equally, only a few seem to be aware of the need to address the issue in their classroom(s) while none of the programs currently adopt a gender perspective while coaching their potential entrepreneurs or when addressing how the entrepreneurial ecosystem functions.
Implications
These findings and perspectives point to the importance of recognizing that a “one size fits all” approach to curricula may not be appropriate, and that gender-sensitive programming, especially related to dealing with these gender stereotypes and gender bias, are needed. This means that in educational development there are opportunities to create better education and create equal opportunities for male and female students.
Value/originality
Women still form the minority amongst the population of (potential) entrepreneurs and find it more difficult to grow their venture due to a range of (institutional) barriers. This study shows that, thus far, EE insufficiently addresses this topic and points to opportunities for interventions for increasing the gender inclusiveness of the entrepreneurial ecosystem, especially for female ones, instilling in them the awareness and knowledge that as a female entrepreneur starting a business isn’t without gendered challenges. This research therefore adds to the body of knowledge on the construction and reconstruction of gender stereotypes and gender bias in the field of EE.
CONFERENTIEBIJDRAGE
Hoe ziet de BIAS bij werkgevers eruit ten aanzien van mensen met een minder geprivilegieerde positie op de Nederlandse Arbeidsmarkt?
ANDERSOORTIG MATERIAAL
DOEL: Deze studie onderzoekt de mogelijke invloed van gender op de historische dynamiek rond verpleegkundig leiderschap. METHODE: Gebruikmakend van een historische onderzoeksbenadering voert deze studie een bronnenanalyse uit met gender als analytische lens, gericht op de ontwikkeling van het verpleegkundig directeurschap in het Sint Radboudziekenhuis vanaf de oprichting van de medische kliniek (1956) tot de uitsluiting van de verpleegkundig directrice uit de directie (1971). RESULTATEN: Er worden zes gendergaps geïdentificeerd, namelijk verschillen in vermeende capaciteiten en kwaliteiten, werk-privébalans, opleiding, salarisstructuur, ondersteuning en gebruik van retoriek. Dit wijst op betrokkenheid van stereotype denkbeelden bij het vormen van de genderasymmetrie binnen het verpleegkundig beroep en de perceptie ervan op de werkplek en daarbuiten. DISCUSSIE: Een geleidelijke uitsluiting van verpleegkundigen op basis van geslacht op strategisch niveau in directies wordt benadrukt. Deze asymmetrie en vooroordelen creëerden een onevenwichtig speelveld, wat de onderhandelingen over de status van het verpleegkundig beroep bemoeilijkte en belemmeringen opwierp voor verpleegkundig leiderschap. CONCLUSIE: Het zichtbaar en bespreekbaar maken van deze vooroordelen kan het bewustzijn vergroten over de wijze waarop historisch gegroeide ideeën en overtuigingen hedendaags verpleegkundig leiderschap beïnvloeden.
ARTIKEL
Affective teacher–child relationships have frequently been investigated in school settings, but less attention has been devoted to these relationships in after-school care. This study explored caregiver- (N = 90) and child-informed reports (N = 90) of the affective caregiver–child relationship (N = 180 dyads) in Dutch after-school care, exploring gender differences at caregiver and child level and the relationship with a gender match between children and caregivers. The caregivers and children reported relatively high levels of closeness and relatively low level of conflict and dependency/autonomy support, irrespective of gender. Multilevel regression analyses revealed that a gender match between child and caregiver was associated with teacher-reported closeness: levels were highest in female-girl dyads and lowest in male-boy dyads. Further, boys indicated the highest levels of autonomy in male-boy dyads, whereas girls indicated the lowest levels in female-girl dyads. Masculinity of staff was associated with more child-reported autonomy support, whereas femininity predicted caregiver-reported closeness in the relationship.
ARTIKEL
This paper conducted a preliminary study of reviewing and exploring bias strategies using a framework of a different discipline: change management. The hypothesis here is: If the major problem of implicit bias strategies is that they do not translate into actual changes in behaviors, then it could be helpful to learn from studies that have contributed to successful change interventions such as reward management, social neuroscience, health behavioral change, and cognitive behavioral therapy. The result of this integrated approach is: (1) current bias strategies can be improved and new ones can be developed with insight from adjunct study fields in change management; (2) it could be more sustainable to invest in a holistic and proactive bias strategy approach that targets the social environment, eliminating the very condition under which biases arise; and (3) while implicit biases are automatic, future studies should invest more on strategies that empower people as “change agents” who can act proactively to regulate the very environment that gives rise to their biased thoughts and behaviors.
ARTIKEL
Albeit the widespread application of recommender systems (RecSys) in our daily lives, rather limited research has been done on quantifying unfairness and biases present in such systems. Prior work largely focuses on determining whether a RecSys is discriminating or not but does not compute the amount of bias present in these systems. Biased recommendations may lead to decisions that can potentially have adverse effects on individuals, sensitive user groups, and society. Hence, it is important to quantify these biases for fair and safe commercial applications of these systems. This paper focuses on quantifying popularity bias that stems directly from the output of RecSys models, leading to over recommendation of popular items that are likely to be misaligned with user preferences. Four metrics to quantify popularity bias in RescSys over time in dynamic setting across different sensitive user groups have been proposed. These metrics have been demonstrated for four collaborative filteri ng based RecSys algorithms trained on two commonly used benchmark datasets in the literature. Results obtained show that the metrics proposed provide a comprehensive understanding of growing disparities in treatment between sensitive groups over time when used conjointly.
HOOFDSTUK
