A recent investigation suggests that AI chatbots may outperform the average human when it comes to creative thinking exercises, specifically in the realm of proposing alternative uses for everyday objects. Nevertheless, it should be noted that even the most adept human participants still exceeded the top-performing chatbots in these particular tasks.
A recent study, recently published in the prestigious journal Scientific Reports, puts forth the idea that large language model (LLM) AI chatbots could potentially excel beyond the capabilities of the typical human in creative assignments, particularly those involving brainstorming alternative applications for common items—a manifestation of what is known as divergent thinking. However, individuals who achieved the highest scores in these tasks still managed to surpass the chatbots’ most outstanding results.
Divergent thinking represents a thought process closely associated with creativity, placing emphasis on the generation of a multitude of distinct ideas or solutions for a given task.
Typically, it is evaluated through the Alternate Uses Task (AUT), wherein participants are tasked with generating as many alternative uses for a commonplace object as possible within a limited timeframe. The responses are assessed across four distinct categories: fluency, flexibility, originality, and elaboration.
In this particular study, Mika Koivisto and Simone Grassini compared the responses of 256 human participants with those of three AI chatbots (ChatGPT3, ChatGPT4, and Copy.Ai) in the context of AUTs concerning four objects—a rope, a box, a pencil, and a candle. The authors gauged the originality of these responses by evaluating them in terms of semantic distance (i.e., how closely related the response was to the object’s conventional use) and creativity.
To quantify semantic distance on a scale from 0 to 2, a computational method was employed, while human evaluators, unaware of the source of the responses, subjectively assessed creativity on a scale from 1 to 5. On average, the responses generated by the chatbots achieved notably higher scores than those provided by humans in both semantic distance (0.95 vs. 0.91) and creativity (2.91 vs. 2.47).
It is worth noting that the human responses exhibited a significantly wider range in both measures—while the minimum scores were notably lower than those of the AI responses, the maximum scores were generally higher. In seven out of eight scoring categories, the best human response outperformed the best response from each chatbot.
These findings suggest that AI chatbots are now capable of generating creative ideas, at a minimum, on par with the average human. However, it should be underscored that this study solely focused on performance in a single task designed for assessing creativity. The authors posit that forthcoming research could delve into how AI can be integrated into the creative process to enhance human performance.
Reference: “Best humans still outperform artificial intelligence in a creative divergent thinking task” by Mika Koivisto and Simone Grassini, 14 September 2023, Scientific Reports.
DOI: 10.1038/s41598-023-40858-3
This research received funding from the University of Bergen.
Table of Contents
Frequently Asked Questions (FAQs) about Creative Thinking Comparison AI vs. Humans
What does the recent study reveal about AI chatbots and creative thinking?
The recent study suggests that AI chatbots may excel in creative thinking tasks, particularly in brainstorming alternative uses for everyday objects. However, it’s important to note that the top-performing human participants still outperformed the best chatbot results in these tasks.
What is divergent thinking, and why is it significant in this study?
Divergent thinking is a type of thought process closely associated with creativity, emphasizing the generation of many different ideas or solutions for a specific task. It is significant in this study because the researchers used it as a measure of creative thinking, particularly in the context of the Alternate Uses Task (AUT).
How was the performance of human participants and AI chatbots compared in the study?
The study compared the responses of 256 human participants with those of three AI chatbots (ChatGPT3, ChatGPT4, and Copy.Ai) in the context of the Alternate Uses Task (AUT). The responses were evaluated based on criteria such as semantic distance (how closely related the response was to the object’s original use) and creativity.
What were the key findings of the study regarding human and AI chatbot performance?
On average, the chatbot-generated responses scored significantly higher than the human responses in terms of both semantic distance and creativity. However, it’s important to note that human responses had a wider range in both measures, with some humans outperforming the chatbots in seven out of eight scoring categories.
What are the implications of these findings for AI and human creativity?
These findings suggest that AI chatbots are now capable of generating creative ideas at least as well as the average human in a specific creative thinking task. However, the study focuses on a single task associated with creativity assessment, and the authors suggest that future research could explore how AI can be integrated into the creative process to enhance human performance.
More about Creative Thinking Comparison AI vs. Humans
- Scientific Reports (Original Research Paper)
- University of Bergen (Funding Source)
5 comments
check out the sci report link for all the details.
chatGPT4 doin’ great in this test, but humans got big range, som super creative!
divergent thinkin important, helps us make new ideas.
wow, this study say AI bots good at thinkin creative, bt humans still bettr in some ways.
glad Uni of Bergen funded this, imp for our future!