What could increasing the model complexity in machine learning cause?

Question

If I increase the complexity of my machine learning model, would training and test errors increase as well?