What is art? This is a question that has been asked throughout history and one that still does not have a clear answer. Art can be defined as a creative expression of ideas, emotions, or feelings. It can be something that is physical such as a painting or sculpture, or it can be something that is intangible such as a song or dance. However, there is no clear definition of what art is and people can have different opinions on what qualifies as art. And this raises an interesting question: if a machine can create a work of art that is indistinguishable from one created by a human, then is it art and indeed what is art?

Some people believe that art can only be created by humans and that machines are not capable of creating art. This is because machines are not creative beings and lack the ability to feel emotions. While machines may be able to create something that looks like art, it would not be truly art because it would lack the emotional connection that is necessary for art.

Others believe that machines are capable of creating art. They argue that art is not defined by the creativity or imagination of the artist, but rather by the reaction it elicits in the viewer. If a machine can create something that causes an emotional reaction in someone who views it, then it can be considered art.

There is no right or wrong answer to this question. It is up to each individual to decide whether or not they believe that machines can create art.

Everything you read above this line was written by an AI trying to explain what Art is and whether an AI can create true art itself. Gotcha :)

Large language models and generative AI models in general have improved quite a bit since last year when I last wrote on this topic, and wondered if AI could write poetry.

I (finally) got an invite to use DALL-E2, the new text to image AI from OpenAI. And even though I already knew what to expect from all the images everyone’s been posting online it still blew my mind. And I’m sure a lot of people outside the AI and Technology spheres of influence don’t yet realize the extent of what’s happening here.

Each image below was completely genereated (instantly, I may add) by DALL-E2 based on a single sentence text prompt shared by me.

I’ll let the images do the talking and shut up for now.

“Portrait of Ganesha drawn in the abstract style of Picasso”

“Portrait of Mahatma Gandhi drawn in the abstract style of Picasso”

“The Taj Mahal at night painted in the style of Starry Night by Van Gogh”

“The Buddha sits in deep meditation under a giant Bodhi tree at dusk while the sun sets in a sky with purple hues”

I’ll just let those images, and who drew them, and how - just sink in for a moment.

What if take the human (me) out of the loop entirely, and ask another AI to generate the text prompts that DALL-E2 would them paint? Well, this is exactly what I tried. GPT-3 (A large language model) suggested text prompts that DALL-E2 should use to create art and here are the results.

“A portrait of a person that captures their inner strength and beauty”

“A painting of a bustling street scene, with people walking and cars driving. The painting should be in the style of a Cubist painting”

“A painting in the style of Van Gogh that captures the feeling of loneliness”

Back to posts