Can computers create great art? Most people struggle to tell the difference between human and digital pictures, and many prefer machine-made paintings.
Test shows we can’t spot AI paintings
Can computers create great art? Most people struggle to tell the difference between human and digital pictures, and many prefer machine-made paintings.
Here is the challenge: 50 pictures in four different styles. Half of them created by artists, the other half generated by AI. Could you tell the difference?
This was the test designed by blogger Scott Alexander. In total, 11,000 people took part online. The average score was just over 60% correct guesses.1
When it came to ImpressionistA 19th Century art movement known for small but visible brush strokes. pictures, the test-takers mistakenly identified every image as human-made - except for the one picture that was not generated by AI. When asked which images they liked the best, even people who claimed to hate AI art preferred it to human work.
According to Alexander, people "can't tell AI art from some of the greatest artists in history".2 But he also thinks that great art is meant to challenge expectations, and that the masterpieces of the 21st Century will be digital creations.
Fans of the technology argue that computers can create images which are "new, valuable, and surprising" - the definition of creativity by cognitiveRelating to the processes of thinking and reasoning. scientist Margaret Boden.
Mathematician Marcus du Sautoy thinks the unpredictable way machines combine different elements is a creative process. "Code exposed to data - existing artworks, for example - is able to learn, mutate and evolve," he argues.3
But many creatives think otherwise. In a popular online essay, the writer Ian Leslie asked whether ChatGPT could compose a prize-winning novel. He concluded that "LLMA large language model (LLM) is a type of AI program that recognises and generates text.s are backward-looking. They are derivativeSomething based on another source or version. , even when creating new combinations."4
People also want a personal relationship with an artist. This explains the popularity of live events, whether poetry readings or music gigs. Though music is easier to stream than ever before, last year music tours reached a record revenue of $9.17bn (£7.21bn) worldwide.5
This could be because audiences respond to artistic intent - the human desire to create and perform. Even Marcus du Sautoy recognises that, "No machine is driven to express itself creatively. It is prompted by the intention of the human."6
But perhaps it is wrong to think of a competition between humans and computers. In the future, creatives could work with AI to make hybrid forms of art. One day, nobody will be able to tell where the human ends and the machine begins.
Can computers create great art?
Yes: The general public cannot tell the difference between human and AI art - and may even prefer the latter. Creativity comes from unexpected combinations, which computers can create as well as people.
No: AI can make convincing copies of artistic styles, but they are still just copies. Machines will not start a new artistic movement by imitating work that already exists.
Or... Artists have historically adopted new technology, from printing presses to cameras and film. In the future, they will find ways to make human work that incorporates AI's abilities.
Keywords
Impressionist - A 19th Century art movement known for small but visible brush strokes.
Cognitive - Relating to the processes of thinking and reasoning.
LLM - A large language model (LLM) is a type of AI program that recognises and generates text.
Derivative - Something based on another source or version.
Test shows we can’t spot AI paintings
Glossary
Impressionist - A 19th Century art movement known for small but visible brush strokes.
Cognitive - Relating to the processes of thinking and reasoning.
LLM - A large language model (LLM) is a type of AI program that recognises and generates text.
Derivative - Something based on another source or version.