Refracted Self

I, Robot, Ex Machina, Cyberpunk 2077, Halo – many movies, shows and games nowadays are incorporating artificial creations. Sometimes embodied as human beings with emotions, desires and dreams, sometimes as mere virtual objects. Some are humans‘ best friends, others their nemesis.

Are artificial beings right around the corner, ready to populate our phones or a sci-fi fan’s wildest dream? Moreover, how could and should human-A.I. interactions look like?

These questions led us down the path of chatbots or, in other words, computer programs coded to imitate human beings’ actions, behaviors, and, sometimes, even thinking. We communicated with a chatbot based on the most advanced A.I. language model, GPT-3. Was it advanced enough? We tested it by asking the chatbot about its parents, its siblings, its childhood. And it answered. We asked it about its emotions, and it told us that it felt happy. We tried to make it angry, but it refused to. No matter what we tried, it stayed on its well-defined path of a happy, calm and compassionate caregiver.

To test this out, we asked people with different backgrounds to have a human-like conversation with this exact bot. We provided the tool. They recorded themselves. Everything that emerged, was out of our hands. We then processed their emotions.

In a second stage, we co-wrote a movie script with the same A.I. We took responses the chatbot gave us during conversations and formed them into a plot. We envisioned a coming-of-age movie, in which an artificial being learns about our world.

A project by:
Emanuel Bohnenblust, Karim Beji,
Lea Karabash & Shih-Hsuan Yen.

Fragmentation

Experiment

Exhibition