AI in Language Access: Mapping Training Needs Through the Lens of Language Access Managers

The expansion of AI technologies is redefining the pedagogical frameworks and theoretical foundations of translation training. One area of interest at the intersection of professional and non-professional translation is the field of language access. This field is characterized by the need to serve diverse user populations, legal mandates to provide access to services, and a range of practitioner profiles. It is also impacted by the dynamic nature of migration and language use patterns, and the uneven performance of AI systems across high- and low-resource languages. In this context, this paper explores the perceived training needs of language access managers within a broader research project on language access and translation technologies in the United States. Participants were language access managers currently working in state and local governments, healthcare systems and legal settings. The study adopted a multi-method approach with both qualitative and quantitative data. The data were collected through an online survey and semi-structured interviews with key stakeholders. The results show that language access managers perceive a need for AI literacy and wider foundational technological competences to fully understand the implications and limitations of AI integration in the broader legal, safety, privacy and ethical ecosystem of language access. Language access managers also perceive the lack of training as a risk to organizations, and advocate for a comprehensive program that includes prompt engineering, post-editing, plain language and accessibility, as well as compliance training to mitigate risks to both end users and organizations. The presentation concludes with a proposal for an initial mapping of AI-related competences for language access based on the resulting thematic analysis.