Emotion recognition is the process of identifying human emotion from facial expressions. This is something that humans do automatically but computational methodologies have also been developing. (for different purposes like surveillance or to drive spending). Remarkable is that all emotion recognition software is built thanks to the elaborate amount of data made by one person: Paul Ekman. The instruction videos of him in which facial muscles are shown are a great inspiration. See http://www.coralievogelaar.com/img/ekman.mov I will build an emotion "chat" bot fed with Ekman's data points. The code of this bot will exponentially raise the possibilities of face movements by using algorithmic image blending and AI as a method, resulting in an infinite amount of remixes and degrees of movements. In this way, even nonexisting emotions emerge. The result will be a duo screen video installation in which 2 faces (live recorded but digital enhanced so looks in between real and artificial) communicate with each other through face movements. The “conversation” will be generated live, but in the code certain rules to make it choreographic and different 'stories' will emerge in the mind of the viewer with a certain uncanny feeling.