There has been much speculation in recent years about both the potential and limitations of art created by artificial intelligence. Technicians and novices alike look at the output of tools like DeepDream and DALL-E, debate their aesthetic merits, and wonder what similar algorithmic-based programs might accomplish in the future. A new documentary Computer slang It reiterates many of the talking points in these debates without adding much to them, but it makes a helpful contribution to the subject: a broad, in-depth look at the actual process of creating art using AI.
In 2016 pop-dance band YACHT – musicians Claire L. Evans and Jonah Bechtolt – decided to use AI to help compose their next album. Directors Reel Roach-Dektor and Sebastian Pardo and their crew followed the band for several years as they pursued this vision, learning about neural networks and strategizing how to extract usable music from the machine. Ultimately, Evans, Bechtolt, and their technical and musical collaborators will notate YACHT’s entire back catalog in MIDI, which will be fed to MusicVAE by Magenta (Google’s open-source project that does machine learning research). With the model “trained” on this information, it can generate its own MIDI snippets based on YACHT’s pre-existing music. The band takes what they can from this output – riffs that interest them, passages that sound good – and then combine the material into various songs. The result is their 2019 album Chain tripping.
disappointed, Computer slang, like most discourses about AI and the arts, operates from a flawed understanding of its own subject matter. Chief among such misunderstandings is whether any of these tools are truly “artificial intelligence.” Algorithm-based programs are nothing more than highly complex aggregators. Although the art they can create looks like human-made art as technology improves, it is not intelligence that is “intelligent”, but a more advanced form of imitation.
The general assumption is that feeding more information into these programs will help them better predict the human mind, as they compute knowledge. But pattern recognition and association are aspects of the sense, not its basis. On their own, these features do not build complex thinking; They create prejudices and stereotypes (as evidenced by AI’s persistent problem with fostering racism). In the beginning Computer slang, an expert warns against anthropomorphizing these tools. Yet even though Evans and Bechtolt jokingly refer to MusicVAE as a “collaborator,” the film still places man and machine in these terms.
It’s a shame because it removes the really interesting aspects of what YACHT does Chain tripping, which explores a novel technique for songwriting. The idea of an “album composed by AI” inspires images of a computer spitting out music, but the process involves a huge amount of human curation and remixing. However, the results sound a lot like YACHT retreading familiar ground… because, really, all the AI can do right now is remix what’s put into it. Still, YACHT’s experiment is worth a try, even if its results are mixed. As a document of that experiment, Computer slang Helps explain the future uses of AI, besides, A lot Go further.
Computer slang Currently in theatres.