While recent progress has been made in several fields of data-intense AI-research, many applications have been shown to be prone to unintendedly reproduce social biases, sexism and stereotyping, including but not exclusive to gender. As more of these design-based, algorithmic or machine learning methodologies, here called adaptive technologies, become embedded in robotics, we see a need for a developed understanding of what role social norms play in social robotics, particularly with regards to fairness. To this end, we (i) we propose a framework for a socio-legal robotics, primarily drawn from Sociology of Law and Gender Studies. This is then (ii) related to already established notions of acceptability and personalisation in social robotics, here with a particular focus on (iii) the interplay between adaptive technologies and social norms. In theorising this interplay for social robotics, we look not only to current statuses of social robots, but draw from identified AI-methods that can be seen to influence robotics in the near future. This theoretical framework, we argue, can help us point to concerns of relevance for questions of fairness in human–robot interaction.