A chain of 10.000 images found through Google’s reverse image search. The visual development is based on the misunderstandings that happen in the AI’s reading of the image material. The AI doesn’t care about scale, emotions or context but is only interested in patterns, colours and correlations.

Footage from Walt Disney's Nature's Half Acre is reworked and given a new voiceover - replacing the paternal voice of Winston Hibler - which also focuses on romantic anthropomorphism. The difference is this voiceover has been generated by a neural network in collaboration with filmmaker Amy Cutler, using an AI which has learned its existence entirely from reading the female protagonist voice in 14 million passages of romance novels.

Using pose detection machine learning models to search over 2 million images from dance television shows, this is a kinetic music video exploring filmic collage techniques, generative textures, audiovisual rhythms, and the universal art of shaking your butt.

The window of an apartment in Rotterdam leads us to the outskirts of an industrial port, a strident and devouring machine that pours out its flames. In front of this infernal and metallic spectacle, the horizon appears obstructed. The steel factories transform the material into smoke, and the boats sail against winds and tides. Above the metal and the buildings that pierce the sky: silence and the flight of and the flight of birds are only a hypothesis.

Using – or rather abusing – some of the latest Artificial Intelligence tools available, Jan sorts, selects and edits film fragments from Amsterdam EYE Film Institute’s unique found footage collection by associating and linking these with the latest online trending topics.

Relationships break down in a dystopian future in this, the first film ever written by an AI.

Can an AI be creative? In early 2021, a machine intelligence spent two days watching several films by the legendary director Jean-Luc Godard. Starting with a blank slate, the AI used these films to learn about our visual world for the first time. Later, a separate AI was used to write the poetry heard in the narration. It is presented verbatim, edited only for length. To complete the experiment a live action narrative, shot in vintage 16mm, was wrapped around this otherwise AI generated film. With machine-generated imagery and narration, the film gives us a glimpse into our world as seen by a new intelligence of our own design.

A double act between drag queen Me The Drag Queen, and a deepfake (AI) clone of Me The Drag Queen. By training a neural network on filmed footage this network learnt to construct a virtual body that can be controlled by feeding it new reference movements.