In principle, the EU GDPR’s Right to Be Forgotten (RTBF) applies to personal data processed by large language models (LLMs) or systems that integrate them, typically AI chatbots. In practice, determining whether something is personal data subject to the RTBF and then ensuring it is forgotten is a non-trivial exercise. Beyond the serious technical challenges, there is no clear consensus among regulators and technical experts as to what is in fact subject to the RTBF. The boundaries defining personal data are fuzzy at best due to re-identification and re-purposing risks. Security threats, complex AI supply chains, and the ability for LLMs to make rapid and revealing inferences further complicate matters.