machines

Our artistic research led us to amass an archive of thousands of recorded worries from people in the US and abroad. Ecology of Worries asks the question of whether we should teach a machine to worry for us. The animation consists of hand drawn critters. Some critters are driven by synthetic worries generated with TextGenRnn recurrent neural network trained on the transcribed worries archive. Other characters are driven to worry by a novel machine learning system called Generative Pretrained Transformer 2 (GPT-2), which was dubbed by some commentators as the AI that was too dangerous to release (but it was released anyway). The creatures’ performance of synthetic worries spans a gradient of intelligibility, reflecting on our deeper collective reality.By characterizing the synthetic worries of various sophistication as variously evolved creatures we aim to engage the empathy of the viewers. It is one thing to experience a text generating neural network failing into mode collapse, which is a state where the system generates the same unchanging output no matter the input (e.g. a string of the same repeating vowel over and over again). It is a whole other thing to watch a mode collapse personified by one of these critters: as we watch the creature struggling to get a word out we can’t stop ourselves from feeling like we should help it finish the sentence. The mode collapse text result of ‘aaa aaaaaaa’ becomes a living wail. The critters in Ecology of Worries appear sentient not because of omniscience a tech evangelist might expect from a digital assistant, but due to their very real flaws. The creatures become uncanny through a juxtaposition of familiar and abstract concerns. The work invites people to watch, listen, and engage with these cute and disturbing beings to make shared concerns—whether serious or hilarious—intimate.

“The Office Diva” is an audio-visual installation; a large scale projection of a computer-controlled character living in a claustrophobic virtual space and compulsively talking. Conceptually, the project is a reproduction and examination of a consciousness ruled by manic-depression. But she is also a machine, and the work plays with the ways in which mad and machinic behavior can manifest in similar ways. Phoebe Sengers argues that the modular design of some intelligent agents makes them hard to understand, they appear to be a schizoid assemblage of random, unmotivated behaviors. Contariwise the computational limitations of other agents have been masked by their insane personalities. Repetitions, lack of affect, inappropriate responses, and non-sequiturs are signs of disturbed people as well as machines. In this project, we deliberately chose a bland machine voice, that speaks the stream-of-consciousness text which is generated, re-ordered and reassembled by a machinic algorithm. But, just as deliberately, we massage the relationship between text and code so that our ”mad” consciousness is not so badly fragmented and fractured as to be indecipherable to a human audience. Over time the bland voice reveals a mad, sad story: a pedestrian story of a receptionist; a perfectionist who works too hard in her small therapy center; a critical observer who sees too much going wrong and strives to fix it; an office Don Quixote tilting at the windmalls of petty inefficiency and corruption; a woman slowing exploding. The voice comes from the psychic emanation of the Office Diva: a larger than life projection of her ego. The graphics represent another synthesis between machine and human; procedural animation creates the flaming or dripping archetype that forms into a dimly human form, that is then reanimated with motion capture. The Diva’s anima swirls chaotically in response to the internal narratives she retells so intently. She drifts in the dimly seen and utterly mundane office environment that has taken on so much overdetermined significance.

(The ELO 2102 Media Art Show.)