Soft robotics deals with the development of soft and flexible robots made of elastic materials. Due to these properties, the robots can be used in sensitive areas or hard-to-reach environments. The currently often experiment-driven design of soft robots requires a high amount of time and resources and is usually application-specific. With the help of AI-assisted design methods, this process is supported. The datasets for training the artificial intelligence are generated by finite element methods (FEM) or the numerical simulations of beam models. Under the specification of external loads and desired deflections as well as other constraints, an optimal design of the robot should be determined. The AI methods are intended to help reduce the number of simulations and experiments and shorten the design time.