Well it kind of looks that way – when you train their network to recognise stuff it can add items to pictures.

Source: Research Blog: Inceptionism: Going Deeper into Neural Networks