laitimes

The AI chatbot ChatGPT showed itself to be the top human thinkers in the Innovative Thinking Test

author:Written by square inch AI

New research reveals AI chatbot ChatGPT shows ability to match top human thinkers in innovative thinking tests**

According to a report by Cai Lian News on July 18, the latest research results of the University of Montana in the United States revealed that the artificial intelligence chatbot ChatGPT performed well in the innovative thinking test, and its ability was comparable to the top 1% of human thinkers.

The University of Montana research team administered the Torrens Innovative Thinking Test on ChatGPT, the most widely accepted measure of innovative thinking. They recorded 8 answers to ChatGPT and compared them with those of 24 University of Montana students.

The AI chatbot ChatGPT showed itself to be the top human thinkers in the Innovative Thinking Test

These answers were later used to compare the scores of the Torrance Innovative Thinking Test for 2,700 college students nationwide. The scoring is conducted by an academic testing service that does not foreknow that it contains answers that contain artificial intelligence.

It is important to emphasize that the Torrance Innovative Thinking Test is copyrighted proprietary material, so ChatGPT cannot "cheat" by accessing relevant information on the Internet or in public databases.

Erik Guzik, a clinical assistant professor at the University of Montana's Business School, is the author of the study, whose research interests include the innovation economy and entrepreneurship. He said the results were unexpected and some of the answers were innovative.

The Torrance Thinking Test consists of both a text assessment and an image assessment, both of which measure divergent thinking, the thought process that produces innovative ideas. The text assessment requires participants to give written responses to pictures or verbal prompts, while image assessments require participants to draw answers.

The AI chatbot ChatGPT showed itself to be the top human thinkers in the Innovative Thinking Test

The study found that ChatGPT ranked in the top 1% for both fluency and originality, but could be improved in terms of flexibility, ability to generate ideas of different types and categories.

Guzik said: "For the first time, we have demonstrated that ChatGPT and GPT-4 can rank in the top 1% in terms of innovation. I was pleased to see that some of my University of Montana students also made it into the top 1%. However, ChatGPT outperformed most college students nationwide. ”

The AI chatbot ChatGPT showed itself to be the top human thinkers in the Innovative Thinking Test

Guzik said they were digging deeper into ChatGPT and found that it had some unexpectedly interesting reactions. These novel and surprising responses prompted them to decide to test their innovative power more deeply.

Read on