In an announcement at the University of Southern Oregon’s Conference on Creativity, OpenAI’s large language model GPT-4 displayed its creative prowess with outstanding results on the Torrance Tests of Creative Thinking (TTCT). Through eight tests, the AI model outperformed humans in the areas of fluency, flexibility, and originality, scoring in the top 1% in all tests for originality – the ability to develop unique ideas beyond conventional norms.
The breakthrough results were achieved by academic Erik Guzik from the University of Montana College of Business, Christian Byrge from Vilnius University Business School, and Christian Gilde from the University of Montana Western. Testing GPT-4 through two variants of the nationally recognized TTCT via ChatGPT, the model produced intellectual questions, revealed hidden causes, speculated potential consequences, and conceived product upgra-des and newly imagined future scenarios.
OpenAI is a research laboratory with the mission of ensuring that artificial general intelligence (AGI) benefits all of humanity. Established in 2015 by Elan Musk, AI researcher Greg Brockman, and technology investor Sam Altman, the laboratory seeks to benefit society by creating powerful general-purpose technologies and techniques. With GPT-4’s impressive showing in creativity tests, the team hopes to find new ways for AI systems to assist humans with focusing on complex and challenging tasks, stimulating idea gener-ation, and evaluating, developing, and testing creative approaches.
It is critical to note that the research team’s findings have huge implications for classrooms, companies, and society as a whole. AI models like GPT-4 can now emulate human creativity to a high extent, driving a re-assessment of how we define, evaluate, and cultivate creative thinking educationally, corporately, and culturally. If AI and humans can recognize the potential for collaboration, the chances of solving complex problems are higher than ever before. Consequently, the research team aims to publish their further studies – including investigation into methods of improving GPT-4’s creative capacity in areas such as business innovation and entrepreneurship – this summer.
