for context on my AI art opinions: i was using what's now called "AI" as part of an installation in 2016 when I made a whole bunch of distorted/deformed/edited artifacts derived from Top Gun and presented them in an installation around a video.
i brought home an ancient Nvidia Tesla C1060 from work since I had a Radeon, heated up my dorm room training an LSTM network on the Top Gun script, and printed the resulting output, which was semantically gibberish but vaguely resembled Top Gun in structure and form. the piece ended up at Bass & Reiner in a pop-up show. nobody noticed what I was doing. I was, as far as I know, one of the very first people making art using generative neural networks as a tool, and nobody noticed.
Along with the script, I had a bunch of 4x6 prints from a Mavigraph dye-sub video printer, some of which I made during the initial presentation of the video, a plaster cast of the video tape from a distorted silicone mold, and all the equipment and materials I used to create the other parts.
the central component of the piece was this version of Top Gun that I painstakingly edited down, frame-by-frame, to remove all narrative content and characters: