This article considers how to define artificial intelligence (AI) systems in employment law. The proposed definitions of AI put forward by the European Commission and the European Parliament in the draft AI ACT are too vague from a labour law perspective. The author hypothesises that research should focus on specific features of the AI system, such as the decisions of automated decision-making systems and their impact on employees' rights and obligations, rather than on a vague notion of AI. The use of appropriate research methods, in particular hermeneutic and axiological methods of legal research, is advisable in the conduct of an AI analysis on the basis of the science of labour law.