Increased demands on efficiency and flexibility with beam operation in recent years required to automate many processes and ensure better modelling of others to allow for deterministic control. Machine learning algorithms have now become standard in the control room in the context of optimisation tasks.
This talk will summarise the latest highlights of some of the ML activities at CERN's accelerators, discuss requirements for potential future machines such as the FCC at CERN and describe the path towards it with the Efficient Particle Accelerators (EPA) project.