Abstract: Fairness has emerged as a key normative touchstone in the discourse around and evaluation of artificial intelligence (AI) systems, with its use stretching across multiple contexts and disciplines. Much has been written on its inter- and intra-disciplinary idiosyncrasies. Whilst the EU embarks upon its ambitious omnibus AI instrument, the AI Act, it should be remembered that AI systems are regulated, albeit less directly, through myriad other legal instruments - some of which use their own conceptualisations of ‘fairness’. This paper takes stock of EU law in data protection, digital markets, content moderation, and the AI Act, in order to analyse the contribution of the term ‘fairness’ in the EU's regulation of AI. It is concluded that the term’s practical utility as a legal term is limited. Fair AI, it is argued, is better placed as a policy ambition, rather than a concrete legal standard to which operators of AI system must adhere.