Back

Snapshot

2022
IAT 222
Group Project

^ Details

View on GitHub

An interactive art project,

For interactive arts course IAT 222, our group of six created Snapshot, a media-translation art installation exploring the difference of creativity and imagination between AI (artificial intelligence) and humans. 

It main attraction is the black box. It symbolizes the obscure nature of the AI's processes, houses the electronics, and divides the space to separate the two participants.

We developed the interaction and artistic decisions as a team and I executed on those decisions. I designed and constructed the black box, wrote all the code to create a seamless experience, and managed all of the equipment required for the final exhibition.

I expanded my skills on project planning, designing, woodworking, and coding (API, I/O, database, synchronization, etc.) through Snapshot.

Concept

AI (artificial intelligence) has been rapidly evolving and has begun appearing in the creative field. Dubbed "Creative AI", these new forms of AI generate new data in text and multimedia mediums. Our goal is to inquire how AI and humans differ when interpreting the same thing.

Execution

One participant is given a 3-6 panel comic. They describe each panel in one sentence. 

A second participant receives the sentences in parallel and creates their own comic in front of a camera, re-enacting one sentence per image.

The sentences are also sent to DreamStudio, an online AI image generator service. The AI never sees the original comic, nor does the second human participant.

The three interpretations and the original comic are displayed publicly for people to discuss, critique, and vote.

Exhibition

It was a rocky start to the exhibition, as teammates were late and a crucial computer component died in transit. Those issues were rectified and the exhibition went smoothly afterwards.

The participants enjoyed the involvement of AI, the craftsmanship of the black box, and the software behind Snapshot.

While the exhibition was a success, there were some things that could have been improved and resolved:
- Create a word filter for banned words.
- Create a more reliable connection between the camera and the software.
- Have a bigger database of comics.
- Automate image capturing for complete software autonomy.

Max Nielsen