Recently, artificial intelligence has been the rave and has advanced in multiple industries. As a fellow content creator, I see A.I being used for video editing, captions, scripts, and visual enhancement. And don’t get me started on ChatGPT! But as A.I progresses, more questions crop up too – how much artificial intelligence should we use? Are we losing our humanity? Personally, the question that I constantly battle with is – are we using artificial intelligence correctly?
The Science Café session in May invited speakers to answer some of these questions, especially those that target the relationship between art and A.I. Here are some rough details from what I gathered at the session. If you find Science Café sessions interesting, do look out for future iterations.
EARNEST WU; Visual Artist – AI, Art and Controversy
To start the session, Earnest introduced the fundamentals of this month’s theme; defining A.I and Art. He reminded us that A.I has been around before in vehicles, social media, and named examples like Siri. They are not limited to the perceived A.I we see in movies like I, Robot. Art, on the other hand, remains to be forms of expression.
Earnest also brought up the question, “What does A.I need to predict?” before going through various models of A.I learning: supervised learning (where a computer is trained on labelled input data), unsupervised learning (input data are not labelled and computer is able to analyze hidden patterns), and reinforcement learning (feedback-based learning with trials and errors).
He also highlighted some controversies that have occurred, such as using A.I to enter and win a traditional photography competition. How can one tell if a photo entry is made genuine or made from A.I? Such questions make us more aware of the use of A.I and its collaboration with art.
IMMANUEL KOH; Assistant Professor, SUTD – AI Aesthetics of Sculpture and Architecture
Professor Koh works with artificial architecture and hypothesizes that there is a hyper surrealist conceptual process when using A.I. There are concepts of surrealism and cubism in training and outputs. Let’s say an AI machine is trained to look at all sorts of chairs, but also to look at all sorts of buildings. It would be possible to find something in between a chair and building.
“A body is a body from any aspect.” Replacing ‘body’ with a noun, let’s say… a building is a building from any aspect. A block with a thousand windows but only one door would still be a building.
From the outputs, Professor Koh also mentions of the aesthetics of defamiliarization and indeterminacy. This means that the output cannot be too perfect nor be too defamiliar as well; there needs to be a sweet spot.
KAPILAN NAIDU; Senior Creative Technologist, Tommy Adjuct Lecturer, Lasalle College of the Arts Media Artist and Curator — Collaborative Creation: Machine Intelligence in Artistic Expression
Interestingly, Kapilan is set to launch an A.I exhibition within Science Centre Singapore for a short period.
His talk covered collaborative creation versus creative collaboration. He notes on the potential to automate the creation process and quicken the ‘boring parts’ to make the overall artistic process easier. Humans should still artistically direct their projects, and only be assisted by machines.
While Kapilan also reminded us how computational art has been around for a while (e.g. speech recognition), he also showcased the extents of artificial intelligence beyond humanity. For example, a computer’s internal logic is different from human discrimination. A computer can think in multidimensional space and organize characteristics, giving it the ability to find something in the middle of everything (like between a chair and a building). It may possess a neural network whereby one input can go to multiple outputs. One of his works involve the ability to track movement around Singapore’s public transport network: where, when, and mode of transport.
Written by Lydia Konig
Photography by Raphael Ng
Check out future sessions of Science Cafe!