These images were created by two artificial neural networks working in tandem. One of those networks, CLIP, has been trained to make connections between English and pictures found on the Internet. The other one, DALL-E dVAE, has been trained to create images based on a large number of parameters. In this toy, CLIP keeps generating new parameters until it starts to perceive that DALL-E is creating images that relate to the title of the movie.
If you want to learn more, please read this great writeup which inspired this toy.