Amputees are beginning to utilize robotic limbs, but such prosthetics typically requires many hours of manual tuning and tweaking to adapt it for the individual. This involves defining the relationship between force and motion of the robotic limbs and assessing an individual’s particular movement style. Research being carried out by North Carolina University, University of North Caroline and Arizona University has shown that this time-consuming trial-and-error process could be significantly reduced using a reinforcement-learning technique which automatically tunes a prosthetic via AI. In this particular case study, a user with a new robotic knee would be able to walk on a flat surface within 10 minutes. Although not at a clinically relevant stage, this cybernetics study is being spearheaded with the view that costly and time-consuming clinic visits can be replaced. Potentially robotic limbs will be able to tune themselves automatically on the go across various types of terrain and to walk up and down steps, but significant amounts of training data will be required.