7 minute read

AI in Creativity: Tool or Collaborator?

Next Article
Editor's Note

Editor's Note

Tommaso de Benedictis on the influence, apprehensions and ethics surrounding the rise of artificial intelligence in the creative industries.

At some point in our lives, whether we like it or not, we have engaged with some form of Artificial Intelligence. The ever-learning technology, emerging straight out of a sci-fi novel, continues to find new applications across most aspects of our daily lives, and looks like it’s here to stay. But what does this mean for the creative industry? Where do we see the relationship of AI and creative work heading in the future? How can we ensure that creative production is protected while at the same time expand creative possibilities and push the boundaries of art, design, and architecture into new domains?

Generative AI platforms such as ChatGPT, Midjourney, and the Generative Fill tool on Photoshop have been gaining popularity not only in the creative industries but across many different sectors. Generative AI can be used to create new content, including videos, audio, code, images, text, simulations, and videos based on given prompts, which can be as simple as one word or as complex as a paragraph. Recent breakthroughs in the field have the potential to drastically change the way we approach content creation (see image below). library due to the lack of transparency around authorship1. A US Court ruled in August 2023 that AI-generated images are not eligible for copyright. This is because “human authorship is a bedrock requirement of copyright” and that copyright has never been granted for work that was absent of “guiding human hand”2. However, we must remember that when photography was popularised it wasn’t allowed to be copyrighted because it was considered simply as a mechanical reproduction3. It took several decades before policies were introduced that recognised the human ‘hand’ or human ‘agency’ behind a photographic image. Interestingly, the term used by the US Supreme Court in the constitution’s copyright clause considers photographs as ‘writings’, in the same way that authors have exclusive rights over their ‘writings’. In the future, the written prompts that we feed Generative AI models for an intended output might also be considered ‘writings’ since they are created by human agency.

While some creatives remain sceptical about the new technology, many of them have embraced it and incorporated into their practice.

Image generated on Fotor.com with the prompt "Portrait of a pug in the style of a Renaissance painting".

There are a plethora of concerns revolving around the use of Generative AI technology: questions around authenticity and authorship, that content created by generative AI models is often derivative and predictable. Potential job losses, data scraping, unethical uses and applications are also of great concern. At the top of the list is the idea these technologies are fed huge amounts of datasets built on the work of countless contributors, more often than not used without their consent. Getty Images, for example, recently announced they will not accept any AI-generated images in their image

The grey areas surrounding copyright and authorship in AI-generated images are something that policymakers today need to contend with hard and fast. The creative industries add £108 billion to the UK economy every year, so if policymakers get it right for creatives, they get it right for the economy. While there are legitimate concerns about the use of Generative AI in the creative industries, there are just as many legitimate benefits. Increased efficiency and cost-effective work, quick decision-making with faster turnarounds and workflows, democratisation of and a springboard for creativity – these are just some of the advantages that Generative AI systems can offer.

Betty Leung, a contemporary London-based artist, is one of the creatives who has actively engaged with AI systems and incorporated them into their artistic practice. Working across different mediums, from sculpture to video art, Betty is most interested in the creative misuses of AI and Machine Learning systems. She engages with these systems in unintended ways, subverting their productivity and recontextualising their output. While she resists making work faster, she hopes this technology can extend her independence and ability to make art for as long as possible, even when her physical dexterity and energy are limited and sight poor(er).

For Betty, these AI systems are sometimes used as a tool in her practice, in the same way a camera is for a photographer, while other times she considers them a ‘collaborator’. More precisely, she sees it as a collaboration with the countless contributors whose work, images, words, and data the algorithm is built upon (the intensive labour that drives these technologies is often obscured, with companies taking credit for the output). The notion that AIgenerated content is made up of hundreds, if not thousands, of layers of information, draws an uncanny parallel to the artist’s own multilayered history: from growing up in Australia of Chinese descent to living and working in London today.

In Betty’s sculptural ‘Interpreter’ series, AIgenerated images alluding to the artist’s own past and identity are printed onto fabric and moulded into amorphous, tubular shapes that are knotted together, free falling to the floor like spilt guts. The AI-generated images that Betty uses for her sculptures are unrecognisable in the same way that Generative AI systems don’t offer a clear-cut view of the data that they’ve been trained on.

When researching material for this article, it was difficult not to dodge all the eschatological ruminations about what this technology holds for us in the future. Betty, however, has provided some real-life suggestions on how to engage with AI ethically today so as not to exploit other’s labour - a sort of guideline that keeps evolving alongside the technology:

• Acknowledge that Generative AI tools are built on datasets containing work of artists and illustrators taken without compensation, consent or knowledge.

• Be transparent about the tools and prompts that are used.

• Don’t imitate or use the names of living or individual artists in your prompt.

• Share and credit your references and inspirations.

• Consider using materials from and contributing to Creative Commons.

Still in its infancy, AI technology is improving at an exponential rate. Fifty years ago, the idea of conversing with a computer named Alexa was pure science fiction. In another fifty years, how are we going to be using AI? We’re at the cusp of something big, and although historically we have always been terrified of the unknown, AI is here to stay, and we should embrace it (I sigh with excitement peppered with disconcertment as I write this). Finding solutions to giving credit to artists whose work has been used for training generative models is key to protecting creator’s work in the future. Art has a history of intention - will AI have its own intention in the future? Will we continue to use AI as a tool, or will it have its own agency and act as a collaborator?

Whichever way it evolves, AI technology opens up huge possibilities for the future of the creative industries.

Tommaso De Benedictis, Associate Curator at Artiq

1. Getty Images Bans Ai-Generated Images Due to Copyright Worries, Artnet News Online, https://www.artnews.com/art-news/news/ getty-images-bans-ai-generated-images-dueto-copyright-1234640201/

2. US Copyright Requirement for ‘Human Authorship’ Enforced in AI Test Case – But That “Bedrock” May be Changing, Stefanie Drawdy, The Institute of Art & Law, https://ial. uk.com/us-copyright-requirement-for-humanauthorship-enforced-in-ai-test-case-but-thatbedrock-may-be-changing/

3. The Evolution of Copyright Law, U.S Copyright Office, https://www.copyright.gov/history/ copyright-exhibit/evolution/

Leading image: Peripheral Interpretations XIII, Betty Leung. Image courtesy of the artist.

This article is from: