A recent study led by Professor Karim Jerbi at the Université de Montréal reveals that advanced generative AI systems can now outperform the average human in certain creativity tests. This extensive research, which compared over 100,000 individuals with current AI models, includes renowned platforms such as GPT-4, Claude, and Gemini, demonstrating a significant turning point in the evaluation of creativity.
The findings suggest that while AI shows strong performance in tasks measuring original thinking and idea generation, the most creative humans, particularly the top 10%, retain a distinct advantage, especially in richer creative endeavors like poetry and storytelling. “While some AI models now outperform the average person, peak creativity remains firmly human,” Jerbi stated, emphasizing that despite AI’s advancements, the highest levels of creativity are still uniquely human.
The study utilized the Divergent Association Task (DAT), a widely recognized psychological tool designed to measure divergent creativity. This involves participants generating a list of nouns that are as semantically unrelated as possible, thereby assessing their ability to produce diverse and original ideas. Although AI models like GPT-4 excelled in this task, researchers noted that the complexity of creative writing—demonstrated through composing haikus, writing movie plot summaries, and crafting short stories—remains a domain where skilled human creators consistently outperform AI.
Jerbi’s research indicates that while AI systems can match human creativity on specific well-defined tasks, they still exhibit limitations. The study’s results reflect a landscape where AI creativity can be manipulated by adjusting technical parameters such as the model’s temperature setting. A higher temperature allows for more varied and less predictable outputs, while a lower setting produces safer responses. This suggests that human interaction and guidance play a crucial role in shaping AI creativity, underscoring the necessity of thoughtful prompting.
The implications of this research extend beyond mere academic curiosity, as it raises questions about the future of creative professions. While concerns about AI potentially replacing human creators are prevalent, Jerbi offers a more nuanced view. “Even though AI can now reach human-level creativity on certain tests, we need to move beyond this misleading sense of competition,” he remarked. Rather than signaling an end to creative careers, the findings suggest that AI can serve as a powerful tool in the creative process, amplifying human imagination and opening new avenues for exploration.
This notion of AI as a creative assistant rather than a replacement highlights a transformative potential for artists, writers, and other creators. By harnessing AI’s capabilities, human creators may find support in expanding their ideas and exploring uncharted creative territories. The research, titled “Divergent creativity in humans and large language models,” is published in the journal Scientific Reports, marking a significant contribution to the ongoing dialogue about the intersection of technology and creativity.
See also
AI-Generated Deception: How Fake Reviews and Algorithms Shape Digital Trust
Microsoft Reveals Copilot Canvas: AI-Powered Whiteboard with Image Generation and Streaming Features
CapCut Unveils AI Video Generation Tools, Enhancing Digital Content Creation for All
Canva Refutes Job Loss Claims Amid Leonardo.AI Integration, Emphasizes Team Stability
Deploy On-Chain AI Agents with Integrated LLMs: A Step-by-Step Guide for Developers
