Female perspectives on algorithmic bias: implications for AI researchers and practitioners
Notice bibliographique
Résumé
Purpose This article explores the use of natural language processing (NLP) techniques and machine learning (ML) models to discover underlying concepts of gender inequality applied to artificial intelligence (AI) technologies in female social media conversations. The first purpose is to characterize female users who use this platform to share content around this area. The second is to identify the most prominent themes among female users’ digital production of gender inequality concepts, applied to AI technologies. Design/methodology/approach Social opinion mining has been applied to historical Twitter data. Data were gathered using a combination of analytical methods such as word clouds, sentiment analyses and clustering. It examines 172,041 tweets worldwide over a limited period of 359 days. Findings Empirical data gathered from interactions of female users in digital dialogues highlight that the most prominent topics of interest are the future of AI technologies and the active role of women to guarantee gender balanced systems. Algorithmic bias impacts female user behaviours in response to injustice and inequality in algorithmic outcomes. They share topics of interest and lead constructive conversations with profiles affiliated with gender or race empowerment associations. Women challenged by stereotypes and prejudices are likely to fund entrepreneurial solutions to create opportunities for change. Research limitations/implications This study does have its limitations, however. First, different keywords are likely to result in a different pool of related research. Moreover, due to the nature of our sample, the largest proportion of posts are from native English speakers, predominantly (88%) from the US, UK, Australia and Canada. This demographic concentration reflects specific social structures and practices that influence gender equity priorities within the sample. These cultural contexts, which often emphasize inclusivity and equity, play a significant role in shaping the discourse around gender issues. These cultural norms, preferences and practices are critical in understanding the individual behaviours, perspectives and priorities expressed in the posts; in other words, it is vital to consider cultural context and economic determinants in an analysis of gender equity discussions. The US, UK, Australia and Canada share a cultural and legal heritage, a common language, values, democracy and the rule of law. Bennett (2007) emphasizes the potential for enhanced cooperation in areas like technology, trade and security, suggesting that the anglosphere’s cultural and institutional commonalities create a natural foundation for a cohesive, influential global network. These shared characteristics further influence the common approaches and perspectives on gender equity in public discourse. Yet findings from Western nations should not be assumed to apply easily to the contexts of other countries. Practical implications From a practical perspective, the results help us understand the role of female influencers and scrutinize public conversations. From a theoretical one, this research upholds the argument that feminist critical thought is indispensable in the development of balanced AI systems. Social implications The results also help us understand the role of female influencers: ordinary individuals often challenged by gender and race discrimination. They request an intersectional, collaborative and pluralistic understanding of gender and race in AI. They act alone and endure the consequences of stigmatized products and services. AI curators should strongly consider advocating for responsible, impartial technologies, recognizing the indispensable role of women. This must consider all stakeholders, including representatives from industry, small and medium-sized enterprises (SMEs), civil society and academia. Originality/value This study aims to fill critical research gaps by addressing the lack of a socio-technical perspective on AI-based decision-making systems, the shortage of empirical studies in the field and the need for a critical analysis using feminist theories. The study offers valuable insights that can guide managerial decision-making for AI researchers and practitioners, providing a comprehensive understanding of the topic through a critical lens.
Récupéré en direct depuis OpenAlex et désinversé. Les résumés ne sont pas conservés dans cette base de données : les index inversés représentent 8,6 Go des 9,3 Go de texte de la base, et le serveur dispose de 13 Go libres.
Comment cette classification a été obtenuedéplier
Prédiction machine sur la base complète
Imitation des enseignantsNi prévalence calibrée, ni vérité terrain. Validation humaine à venir. Le volet Gemma est une étiquette directe du modèle pour chaque travail de la base, lue sur la notice réduite au titre. Le volet Codex est un classifieur appris des 10 348 étiquettes directes de Codex et calibré sur les taux pondérés de l'échantillon; les champs sans appui suffisant ne portent aucun appel Codex. Le mode candidate est l'union des deux volets; le consensus est leur intersection. Ces sorties portent le statut machine_predicted_unvalidated et ne sont pas des étiquettes humaines.
Scores du classifieur distillé par catégorie (deux têtes)
| Catégorie | Codex | Gemma |
|---|---|---|
| Métarecherche | 0,045 | 0,078 |
| Méta-épidémiologie (sens strict) | 0,001 | 0,000 |
| Méta-épidémiologie (sens large) | 0,001 | 0,000 |
| Bibliométrie | 0,004 | 0,003 |
| Études des sciences et des technologies | 0,007 | 0,027 |
| Communication savante | 0,012 | 0,016 |
| Science ouverte | 0,002 | 0,007 |
| Intégrité de la recherche | 0,003 | 0,005 |
| Charge utile insuffisante (le modèle a refusé de juger) | 0,014 | 0,002 |
Scores machine (provisoires)
Les deux têtes enseignantes du modèle étudiant, lues sur ce travail. Un score ordonne la base pour la relecture; il n'affirme jamais une catégorie, et le statut de validation accompagne chaque rangée tel quel.
Scores de référence d'un modèle non mature (critères de maturité non atteints, 7 itérations). Un score ordonne; il n'affirme jamais une catégorie.
score_only:v0-immature-baseline · tel quel depuis la passe de notation : score_only signifie que le nombre peut ordonner les travaux, et qu'aucune étiquette de catégorie n'en découleClassification
machine, non validéePrédiction automatique; un appel candidat d’une seule source (Gemma direct ou Codex distillé), pas un consensus.
Le détail, modèle par modèle et score par score, se trouve en fin de page sous « Comment cette classification a été obtenue ».