What we can mean when we talk about personal data is constantly changing with the ongoing development of information and communication technology. In both European and Swiss law, the interpretation of this term is of central importance for data-related personal data protection, as only personal data opens up the scope of the GDPR and Swiss data protection laws. With the advent of learning AI in the form of large language models (LLM) in the mainstream of society, the question arises as to whether and, if so, how the concept of personal data and thus the regulatory object of data protection law is changing as a result of the shift from evidence-based or ‘causal’ data to statistics-based or ‘stochastic’ data. It appears that the answer is related to the semantic interpretability of the link between persons and data.