Scientific and technical activities on JET focus on the issues likely to affect the ITER design and operation. Our understanding of the ITER reference mode of operation, the ELMy H-mode, has progressed significantly. The extrapolation of ELM size to ITER has been re-evaluated. Neoclassical tearing modes have been shown to be meta-stable in JET, and their beta limits can be raised by destabilization (modification) of sawteeth by ion cyclotron radio frequency heating (ICRH). Alpha simulation experiments with ICRH accelerated injected 4He beam ions provide a new tool for fast particle and magnetohydrodynamic studies, with up to 80-90% of plasma heating by fast 4He ions. With or without impurity seeding, a quasi-steady-state high confinement (H98 = 1), high density (ne/nGW = 0.9-1) and high β (βN = 2) ELMy H-mode has been achieved by operating near the ITER triangularity (δ ∼ 0.40-0.5) and safety factor (q95 ∼ 3), at Zeff ∼ 1.5-2. In advanced tokamak (AT) scenarios, internal transport barriers (ITBs) are now characterized in real time with a new criterion, ρT*. Tailoring of the current profile with lower hybrid current drive provides reliable access to a variety of q profiles, lowering access power for barrier formation. Rational q surfaces appear to be associated with ITB formation. Alfvén cascades were observed in reversed shear plasmas, providing identification of q profile evolution. Plasmas with 'current holes' were observed and modelled. Transient high confinement AT regimes with H89 = 3.3, βN = 2.4 and ITER-relevant q < 5 were achieved with reversed magnetic shear. Quasi-stationary ITBs are developed with full non-inductive current drive, including ∼ 50% bootstrap current. A record duration of ITBs was achieved, up to 11 s, approaching the resistive time. For the first time, pressure and current profiles of AT regimes are controlled by a real-time feedback system, in separate experiments. Erosion and co-deposition studies with a quartz micro-balance show reduced co-deposition. Measured divertor thermal loads during disruptions in JET could modify ITER assumptions.