What about the Artists? An exploration into how AI is changing the art sector as we know it.
The rapid advancement of technology in recent years has brought about a significant change in various industries, and the proliferation of AI models used to generate art got me thinking — what about the artists? Can we consider the builders of these algorithms new age art contemporaries, or are they just thieves? With a landmark piece of AI-generated art selling for $432,500 in 2018 after a valuation of just $10,000, and galleries embracing art created through AI programmes, it came to me that we should probably be questioning the opportunities and challenges that face workers in this sector, from artists to photographers to programmers.
One of the most significant opportunities AI art software brings to the art and photography sector is increased efficiency and accessibility. When scrolling through TikTok a couple of weeks ago, I came across a trend of users paying for AI-generated headshots through the use of websites such as SuitUp and TryItOnAI. Customers can create high-quality portraits in a fraction of the time it would take for them to hire a photographer, get ready and get the photos taken and edited. The cost for this process could cost between £200-£450 when hair and makeup are taken into account, and the process can be long, and uncomfortable for some.
I discussed with Delilah about how she felt using this AI programme, and the answers may have photographers wanting to cover their ears. For the low price of $17 (or around £14), Lila said she was able to have 100 different options generated and she was highly impressed with the results! She went on to share this ‘hack’ with her fellow co-workers who were also due an updated profile picture on their Linkedin and said ‘half my team ended up using it, so that’s 11 people right there.’ When asked if she would consider using AI for other portraits in the future, she said she had already invested money into an AI-generated artwork piece for her living room, made using OpenAI’s DALLE programme.
It's clear to see then, that AI art does have demand, and whilst that demand may not be valued on par with work done by traditional artists, there is definitely a market opening for quickly generated images for consumers and fine art for collectors. The opportunities for revenue generation in this way are quite therefore incredibly high, as once an algorithm is written, it can create the art passively and the only labour required is to keep the programme up and running as the AI is self-learning and always developing from new inputs. The DALLE neural network for example is trained using a massive collection of images, and can generate new pieces from textual prompts.
I think it’s fair to ask the question, where are images that OpenAI are training DALLE coming from then? I utilised OpenAI’s chatbot ‘ChatGPT’ about where they get their images from, but the answer it gave me was not very illuminating. It replied “It is possible that OpenAI obtained a license or permission to use the images for the purpose of training their AI algorithm. However, it is also possible that they used images without permission, relying on the fair use doctrine for research and development purposes.” It seemed that the fact OpenAI wasn’t transparent about the permissions they had for the images DALLE was trained on when asked makes me suspicious about the legal implications.
Copyright law surrounding AI-generated art is a relatively new and complex area. One notable case related to copyright law and AI-generated art is the dispute over the ownership of the artwork Portrait of Edmond de Belamy made by art collective Obvious, using an algorithm to generate the image. In this case, the code was taken from GitHub, uploaded by artist and programmer Robbie Barrat, and he has accused them of failing to credit his work. The lack of precedent for legal cases involving AI-generated artworks leaves a major challenge to artists that use AI, programmers who feel the work generated on their platforms belong to them, and further, artists who feel their art has been misused when training an AI algorithm to produce images.
Another challenge posed by AI art software is the potential for job displacement. As technology becomes more advanced, it could be possible for AI algorithms to create art and photography at a level that is comparable, or even superior to human-created work in the near future (today is the worst that AI will ever be again, remember). This could lead to a reduction in demand for human artists and photographers, particularly in areas such as commercial photography and portraiture. To rise to this challenge, photographers and artists need to begin to think about what their service includes that a client couldn’t get from artificial intelligence.
For example, when it comes to Delilah’s photos there were quite a few headshots which came out distorted, with a common feature being teeth fused to the lips in open-mouthed photos, or distortions causing the eyes to be glassy or point in different directions. Whilst the AI programmes give customers 100 images to choose from to mitigate this issue, when it comes to a professionally shot photo this isn't an issue a photographer would run into. Furthermore, quite a few reports of customers using the apps reported that the headshots generated would be ‘convincing to someone who doesn’t know me,’ and this got me thinking about Masahiro Mori’s ‘uncanny valley’ phenomenon which refers to a point where the resemblance to a human is so close but not quite there, that it triggers a feeling of revulsion in humans. It's certain that some of the photos developed by AI trigger that feeling, at least for me.
The use of AI art software also raises ethical concerns, particularly in areas such as portraiture. AI headshot generators like SuitUp use facial recognition technology to create portraits, and as Michael discussed within the ethics content, this raises questions about privacy and consent. There are also concerns about the potential for AI algorithms to perpetuate or even amplify biases, particularly in areas such as race and gender.
The introduction of AI into the creative sector is still an extremely new event and we can’t predict the future when it comes to consumer habits, but it seems very likely that AI artwork generators are here to stay. If algorithms are made conjunctly with artists, they might be able to hang on through these technological developments, instead of being left in the dust. And still, the AI programmes will always need new data inputs.
