Skip main navigation

Meet the artist: Terence Broad

An interview with artist Terence Broad where his discusses his approach and explains some of his seminal work.
Artist Terence Broad

You’ve just read about Terence Broad and some of his seminal work. We interviewed Terence to find out about his work as a creative AI artist.

He explains his approach to the intersection of creativity and technology, and using machine learning models and algorithms as artistic materials.

How would you introduce yourself? Tell us what you do, and why it’s important.

I am both an artist and a researcher completing my PhD at Goldsmiths. In my research I develop tools with machine learning that can be used by artists and designers. I also make my artworks with the methods that I develop in my research.

Can you talk us through some examples of your work as an artist and researcher?

I made a series of artworks entitled (un)stable equilibrium where I trained GANs – a type of generative neural network without any data. I was motivated in trying to find a way of using AI to generate something truly novel, rather than learning to recreate images from photographs or existing paintings. Ironically the results look remarkably like paintings by Mark Rothko, but this was not the intended result when I set out to make this work.

Another set of artworks I made recently that has come out of my research was a series titled Being Foiled which once again came out of my line of research investigating how to get generative machine learning algorithms to generate images that were completely novel, and not work that had been seen before. To make these works I took a GAN that had already been trained to generate realistic looking images of faces, and optimised it towards producing images the algorithm thought looked fake rather than what looked real. The system ends up producing images of faces that are very uncanny and unsettling (which you can read more about in my blog.

What are the main problems that this work in particular tries to address/explore?

I am interested in using these generative neural network systems as raw materials that can be played around with and explored in interesting ways. These are incredibly powerful systems and there is a huge amount of potential for creating new aesthetic possibilities. I would describe my artistic practice as interrogating new ways these tools can be used and manipulated.

How did you get involved in creative AI?
What led you to explore these technologies in this way?

I started this work on my Masters in Creative Computing and I am now continuing them on my PhD. Previous to my technical education I was an art school dropout. I left a degree in Sculpture because I was fascinated by the potential for computers to be used in image making and image processing and I decided I needed to learn programming to pursue that properly.

What are some of the challenges in this field?

Some of these systems require large expensive GPU computer systems to train and to run. Lowering the computational power needed to run these systems, and letting people run them on platforms that are more user friendly than Linux would help lower the barrier to entry to new people who want to experiment with this stuff.

Could you tell us about some of the recent developments in this field?
What should we be looking out for in the next year or next 5 years?

Some of the most exciting developments have been with language models (in particular with recent advances with a model called Transformers) which are capable of producing incredibly lucid text. In my field of image making, StyleGAN and StyleGAN2 are the most impressive methods for producing realistic images.

I follow lots of very exciting people who are now capable of training their own models using RunwayML, an accessible easy to use method for training your own CreativeAI models. I think as these tools get even more accessible there is going to be lots of new people producing all sorts of exciting things.

What are the skills you need to do your job, and work in Creative AI?
How can someone get these skills?

Patience! It can be incredibly frustrating getting these systems set up, and it can be difficult to debug when it goes wrong. A great deal of patience is required to keep going when things aren’t going well.

What other advice do you have for our learners who’d like to do what you do?

Try and do things other people aren’t doing, and don’t get disheartened if you find out people have already done your ideas, or you don’t think your work is as cool as other peoples. It takes time to find a niche, but there is still a huge amount of potential for people coming in with fresh ideas and perspectives to contribute to the field.

This article is from the free online

Introduction to Creative AI

Created by
FutureLearn - Learning For Life

Reach your personal and professional goals

Unlock access to hundreds of expert online courses and degrees from top universities and educators to gain accredited qualifications and professional CV-building certificates.

Join over 18 million learners to launch, switch or build upon your career, all at your own pace, across a wide range of topic areas.

Start Learning now