Artificial Intelligence gains another win over humans, as a new study suggests it can have \u2018creative thought.\u2019<\/span><\/p>\n In a <\/span>study<\/span><\/a> conducted by PhD students at the University of Arkansas, 151 people were pitted against ChatGPT-4 with the aim of measuring divergent thinking.<\/span><\/p>\n This is often thought to be a sign of creativity and imagination, as it involves a thought process that is used to generate ideas and explore possible solutions to a concrete or abstract issue.<\/span><\/p>\n In the study, the multimodal language model created by <\/span>OpenAI<\/span><\/a> was found to provide more \u2018original and elaborate answers than the human participants.\u2019<\/span><\/p>\n The first take involved people coming up with creative uses for everyday objects (like a fork or rope). The second called the Consequences Task, asked people (and their AI counterparts) to imagine the outcomes of a made-up situation. The third, named the Divergent Associations Task, required 10 nouns that are as \u2018semantically distant as possible.\u2019<\/span><\/p>\n Researchers then looked at the length of response and semantic difference between words. The <\/span>ChatGPT-4<\/span><\/a> model was found to be \u2018more original and elaborate than humans on each of the divergent thinking tasks, even when controlling for fluency of responses.\u2019<\/span><\/p>\n While the full insight into AI\u2019s creativity hasn\u2019t yet been extensively looked at, this study marks one of the first in the exploration of its capabilities.<\/span><\/p>\n Conducted by Kent F. Hubert, Kim N. Awa, and Darya L. Zabelina, the students have explained that some caution should be exercised.<\/span><\/p>\nResearchers have some \u2018caveats\u2019<\/span><\/h2>\n
![\"\"](\"https:\/\/readwrite.com\/wp-content\/uploads\/2024\/02\/avatar_user_26583_1708362924-125x125.jpeg\"){kind=avatar}