Divergent thinking is related to creativity and is related to the generation of different alternatives or solutions for a given task. The usual way to measure it is to invent the different uses that a current object can have for a limited time. And they've done it just to compare the divergent thinking of three artificial intelligence text creators (chatbots) with 256 volunteers. The result is that chatka boots are better than humans. However, the best volunteers were better than the best chats.
The researchers delivered four objects to the three chatrooms (ChatGPT3, ChatGPT4 and Copy.Ai) and the 256 volunteers to invent the uses: a rope, a box, a pencil and a candle. To measure the originality of the responses, researchers have taken into account creativity (1 to 5) and the semantic distance between the response given and the original use (0 to 2). On average, the responses of chatbots have been better than those of humans in both parameters: creativity 2.91, human 2.47, and semantic distance 0.95 and 0.91, respectively.
The difference between the highest and lowest scores of volunteers was much greater than that of the chatrooms. And the results of the best humans were pretty much better than the chotoads.
The work has been attended by researchers from Finland and Norway and has been published in the journal Scientific Reports. Within the limits, the homogeneity of the volunteer group (all young or middle-aged adults from western culture) and the inability to compare the processes (among other things, because chats are “black boxes”) are mentioned. However, they propose to investigate how to integrate chatbots into future creativity processes.