The Struggle With AI Generated Art

A new trend has been on the rise where ordinary people have been using AI tools to “generate” artwork. An AI art generator is a tool that can generate or create artwork or an image by taking a couple of words or a sentence as input. Among these tools, the most popular ones are named Dall-E by Open AI and Stable Diffusion by Stability AI. These tools use state-of-the-art neural networks, popularly known as GANs (generative adversarial networks). It’s a collection of two systems, where one learns from a dataset of images by recognizing a similar pattern and memorizing it and the other will try to reproduce the learned pattern and create images based on a provided description. A sample input description would be “An image of an astronaut eating cereal on Mars”.

The AI will try to create an image of this complexity and provide more than one output of different art forms or styles which can be further modified or customized or re-fed back into the algorithm for better suggestions. While this technique of creating art is technologically advanced, it poses potential ethical challenges.

Struggle From Artists

Artists create their art, hone their skills over time, and adopt a style or a system that makes them unique. That’s the selling point of an artist, how unique and creative they can be. An AI system, however, uses artwork or images from various artists and will learn their patterns and skills, and its artistic instincts will be significantly better because a computer is in many ways better than humans. For example, these AI systems can also take an input prompt like “generate an image of a person riding a horse near the river, in the style of the famous painter Van Gogh”.

An AI system will understand this input and would have already recognized patterns from Vincent van Gogh and will produce shockingly accurate results about the art style of the said painter. While these tools cannot be customized heavily as per the requirement, they perform “just enough” to automate and replace human artists or designers. For corporates, this could save a ton of time and money. The ethical challenge here is that AI systems not only steal the art form or style but are also doing a better job than a human could ever do. Artists start early and spend their whole lives honing their skills and being an artist, itself is hard enough because of the fierce competition and uncertainty of freelance job markets. Now, they face an all-powerful smart AI that could potentially render them worthless. Artists also have an ethical means of maintaining intellectual property rights using signatures, watermarks, or digital signatures (in the case of digital artwork). An AI-generated artwork has no trace left behind because of the nature of how these machine-learning algorithms behave, it can be hard to see the inner workings of these systems.

Developer's Perspective

The creators of these tools argue that the nature of AI systems enables them to be this smart and that not everything they do could be a theft of individual property. They also say that these AI systems as merely tools, with the right mindset, can be used by artists themselves to create art in a unique never-before-seen perspective. They can also use it to improve their creative thinking like AI-enabled writing assistants helping writers improve. These tools will also enable non-artists to create high-quality artwork and even tinker around their thoughts limited to their artistic styles.

Conclusion

Human artists are hoping that there will be laws that will protect their creative pursuit and let technology be developed in a non-invasive way. They also endorse the emerging creator economy that enables small and remote artists to understand something clearly at last. On the other hand, developers are arguing that AI art creation is not a machine creating art but it’s just following a bunch of instructions and is not trying to communicate anything. Hopefully, in the future we see both artists and developers pursue their creative work without having to go against each other, instead, they can use automated tools to be more creative and innovate things in ways that were not possible before.