The Council of Europe (2022) positions plurilingual and intercultural education together as fundamentally interconnected, and The Common European Framework of Reference (CEFR) materialises this vision through the concept of mediation. Yet in UK higher education, Intercultural Communicative Competence (ICC) remains peripheral across language curricula. This essay interrogates this gap and asks whether artificial intelligence (AI) can act as a practical bridge. Drawing on a mixed-methods survey of 60 UK Higher Education language teachers, it maps roles, usage patterns, perceived benefits, and risks of AI for language teaching and ICC. The analysis of data shows sporadic adoption, with frequent users identified as the category leveraging AI more often for interactive, practice-based activities (e.g. simulated intercultural role-plays). Across groups, concerns about bias, cultural misrepresentation, privacy, academic integrity, uneven access, and limited institutional support for embedding AI emerge. The essay situates these insights within the CEFR notion of mediation and presents a concrete case of supervised AI-supported ICC activities across seven languages at beginner and intermediate levels. Although student feedback is preliminary, it suggests joint gains in language proficiency, intercultural awareness, as well as on digital literacy. The conclusion argues for moving beyond scattered experiments towards sustained and ethical integration: aligning learning outcomes and assessment with ICC; building iterative, intergenerational CPD; embedding critical AI literacy; and prioritising less commonly taught languages in research to truly gauge both AI potential and limits to facilitate a move of ICC instruction from margins to mainstream.