Linguistic and cultural bias in AI. Implications and strategies for teacher education
Artificial Intelligence (AI) is transforming global communication and education, but its inherent linguistic and cultural biases present challenges that must be addressed in pre-service teacher education. Countries such as China and Japan are developing their own AI models, motivated by concerns over English-centric systems that fail to represent their languages and cultures adequately. Linguistic biases, particularly in English AI systems, disproportionately affect users of non-standard dialects like African American Vernacular English (AAVE), as studies reveal AI models may associate these dialects with negative outcomes, such as discriminatory treatment in hiring processes, denial of housing or rental opportunities, and other forms of systemic inequity like harsher sentencing recommendations. This chapter examines these biases, contextualizing them within the field of education. The objective is to explore how future educators can critically analyse and address AI bias and foster a responsible use of Generative AI systems. Strategies include integrating activities that highlight linguistic diversity and fostering culturally responsive teaching practices. Ultimately, this chapter emphasises the importance of preparing educators to mitigate bias and to equip students with tools and techniques for navigating an increasingly AI-driven world, promoting equity and inclusivity.
Also available via: European Organization for Nuclear Research