Trained machine learning models allow for computationally expensive processes to be done quickly. In my work, I explore whether we can use machine learning to create machine learning emulators that predict atmospheric responses to various forcings. My first project used data from a dry dynamical core to predict the jet stream response to temperature perturbations. Following the success of that work, I am now training machine learning emulators on more complex data while also exploring what useful information machine learning emulators could provide to climate scientists.