People say they prefer a short story written by a human over one composed by artificial intelligence, yet most still invest the same amount of time and money reading both stories regardless of whether it is labeled as AI-generated.
That was the main finding of a study we conducted recently to test whether this preference of humans over AI in creative works actually translates into consumer behavior.
Amid the coming avalanche of AI-generated work, it is a question of real livelihoods for the millions of people worldwide employed in creative industries.
To investigate, we asked OpenAI’s ChatGPT 4 to generate a short story in the style of the critically acclaimed fiction author Jason Brown.
We then recruited a nationally representative sample of over 650 people and offered participants US$3.50 to read and assess the AI-generated story.
Crucially, only half the participants were told that the story was written by AI, while the other half was misled into believing it was the work of Jason Brown.
After reading the first half of the AI-generated story, participants were asked to rate the quality of the work along various dimensions, such as whether they found it predictable, emotionally engaging, evocative and so on.
We also measured participants’ willingness to pay in order to read to the end of the story in two ways: how much of their study compensation they’d be willing to give up, and how much time they’d agree to spend transcribing some text we gave them.
So, were there differences between the two groups?
The short answer: yes.
But a closer analysis reveals some startling results.
