With the recent advances in artificial intelligence (AI), a new tool for art generation has been born. This tool, called AI art, allows users to type in writing prompts and generate unique pieces of art based on those prompts. Some people believe that this tool will destroy art as we know it, while others believe that it will be a savior for the art community.
There are two main AI art technologies that have gained attention in recent months: MidJourney and OpenAI’s DALL-E2. Both tools take different approaches to art generation, with MidJourney being more artistic and painterly, and DALL-E2 being more photorealistic.
Several other AI art technologies have also been developed by other companies, including Google Brain’s Imagen, Microsoft’s NUWA-Infinity, Stable Diffusion, and Craiyon (using DALL-E Mini). Adobe is also developing its own AI art software.
With the influx of investment into AI art creation, many people are worried that human-created art and procedurally generated art will soon become indistinguishable. In fact many well-known artists are now finding that their fans are using AIs to recreate their styles, producing new works with the distinctive look of these artists, but with subject matter than the fans select.
While in most cases these AI created works are easily distinguishable from the ‘real thing’, over time this gap is likely to fall dramatically – with the end result being that anyone can replicate any artist’s style to produce new works, without the artist receiving any compensation.
This is a challenging situation for artists focused on commercial success, as ‘counterfeit’ art, on original topics but using their distinctive styles, is already cheap to produce and easy to distribute. Governments have yet to take a stand to support the art world to draw a line between human and AI art, and their ability to do so is likely to be limited.
So will the art world survive the coming flood of AI masterpieces? Will humans remain at the centre of creativity and profitability across the arts?
Unfortunately the prospects currently look dim for this. While humans can still enjoy creating art, and there’s inherent value in artworks produced by humans for humans, this value may not sustain the same price tags and commercial opportunities for artists as existed before AI.
Concerted effort would be required to design systems to accurately label human and AI art – likely involving AIs themselves – and except in the case of the most expensive artwork, people decorating their homes and businesses may not care whether an artwork was ‘real’ (by a human) or ‘fake’ (by a machine) and instead focus on the impact the art has on them as an individual and how well it suits their decor and style.
AI art may produce a renaissance of sorts, with AI capable of developing works in a range of styles never attempted by humans, and at a speed humans cannot match. This may also create new opportunities for human artists, however is unlikely to trigger a human renaissance in its wake.