Images made by artificial intelligence (AI) are becoming more common, showing up in ads and on Instagram. As image generators like Dall-E 2 become more powerful, easy to use, and popular, agencies may run into technical and practical problems when they try to use them. Most intriguing, though, are the intellectual, theoretical, and ethical problems that marketers must consider.
So, what exactly is creative thinking?
In Marcus du Sautoy’s definition of creativity in his book The Creativity Code: AI art is the ability to make something new, unexpected, and valuable. An AI photocopier or static machine could be made very quickly, but an AI that could generate something novel, unexpected, and useful would be a significant accomplishment.
Can the creator also be the judge?
AI models with a wide range of uses, like Dall-E 2, can analyze a wide range of visual data. When you need to focus on a certain subject to connect with a certain group of people, the models fail miserably.
These pictures will be the only ones the AI has ever seen, making them unique among the works of human artists. Building a compelling dataset is important since the AI will only ever know what is in the dataset, and that data will have a significant impact on the AI’s output.
To train our AI to reproduce the spirit of street art, with its vivid colors, angular fonts, and graphic shapes, we collected a collection of street art in Prague that spans several decades, forms, themes, and styles. Our AI performed a fantastic job of replicating those nuances once it was taught.
How close to human is AI?
It’s a huge task to develop an AI artist that can compete with human creativity. Technically speaking, an AI has the capacity for infinite creativity; it can produce an infinite number of pictures. The challenge that many advertising companies will confront is how to choose the most effective pictures from a large pool of potential options generated by AI to use in their campaigns.
The subjective nature of the project’s conclusion echoed our initial debates about creativity, even if it was aided by the technology we developed. Ultimately, this was the project’s most valuable insight: AIs are the result of a distinctly human process. No matter how sophisticated an AI is or what field it operates in, it will always lack the creativity, common experience, and prejudices that humans possess.
An actual group of humans is always working behind the scenes of every high-quality equipment. When comparing creative time to technical time, creative time always wins. Although there is a lot of anxiety about automation, I think we should be bold enough to use cutting-edge software to address intractable issues. However, it’s important to keep in mind that AI is still simply a piece of software.