Can generative artificial intelligence systems like ChatGPT genuinely create original ideas, or are they merely sophisticated mimics of existing human ingenuity? This pivotal question has been at the forefront of technological and philosophical debate for years, and a new, groundbreaking study offers the most comprehensive answer to date. Led by Professor Karim Jerbi from the Department of Psychology at the Université de Montréal, with the significant participation of renowned AI pioneer Yoshua Bengio, the research represents the largest direct comparison ever conducted between human creativity and the capabilities of large language models (LLMs), conducted at an unprecedented scale. Published in the esteemed Scientific Reports (a Nature Portfolio journal) on January 21, 2026, the findings signal a profound shift in our understanding of machine intelligence, revealing that AI has not only caught up to, but in some aspects, surpassed the average human on specific creativity measures. However, the study also unequivocally affirms that the most creative individuals continue to hold a distinct and consistent advantage over even the most advanced AI models, offering a nuanced perspective on the future of creativity in an AI-augmented world.
A New Benchmark: AI Reaches Average Human Creativity Levels
The research meticulously evaluated several leading large language models, including prominent systems such as OpenAI’s ChatGPT (specifically GPT-4), Anthropic’s Claude, and Google’s Gemini, among others. Their performance was rigorously benchmarked against the responses of over 100,000 human participants drawn from diverse demographics. The results presented a clear and undeniable turning point in the trajectory of AI development: certain sophisticated AI systems, particularly GPT-4, demonstrated the ability to exceed average human scores on tasks specifically designed to measure divergent linguistic creativity. This achievement marks a significant milestone, shifting the perception of AI from a tool for computation and data processing to one capable of generating novel and varied ideas, a hallmark of creative thought.
Professor Karim Jerbi articulated the gravity of these findings, stating, "Our study shows that some AI systems based on large language models can now outperform average human creativity on well-defined tasks. This result may be surprising—even unsettling—but our study also highlights an equally important observation: even the best AI systems still fall short of the levels reached by the most creative humans." This dual insight forms the core of the study’s contribution, offering both a testament to AI’s rapid advancement and a reassurance of unique human capabilities.
Further in-depth analysis conducted by the study’s co-first authors, postdoctoral researcher Antoine Bellemare-Pépin from the Université de Montréal and PhD candidate François Lespinasse from Concordia University, illuminated a striking pattern. While the average person’s creative output could now be matched or even eclipsed by certain AI models, the apex of creativity—the highest levels of originality and divergence—remained firmly within the human domain. This distinction is crucial for understanding the practical implications of AI in creative fields. When researchers focused on the most creative half of the human participants, their average scores consistently surpassed those of every AI model tested. This gap widened even further and became profoundly pronounced when examining the top 10 percent of the most creative individuals, underscoring a qualitative difference that current AI models have yet to replicate.
Professor Jerbi, who also serves as an associate professor at Mila (the Quebec AI Institute), emphasized the methodological rigor behind these conclusions. "We developed a rigorous framework that allows us to compare human and AI creativity using the same tools, based on data from more than 100,000 participants, in collaboration with Jay Olson from the University of Toronto," he explained. This commitment to a standardized, objective comparison across both biological and artificial intelligence systems ensures the robustness and reliability of the study’s conclusions.
The Methodology: How Scientists Measured Creativity in Humans and AI
To ensure a fair and equitable evaluation of creativity across such disparate entities as humans and machines, the research team employed a multi-faceted approach. The primary instrument for measurement was the Divergent Association Task (DAT), a widely recognized and validated psychological test specifically designed to assess divergent creativity. Divergent creativity is defined as the ability to generate a broad range of unique, varied, and original ideas from a single prompt or starting point, a key component of creative problem-solving and innovation.
The DAT, originally developed by study co-author Jay Olson, presents a seemingly simple yet profoundly insightful challenge. Participants, whether human or AI, are instructed to list ten words that are as semantically unrelated in meaning as possible. The power of the DAT lies in its simplicity and its capacity to reveal the breadth and originality of an individual’s associative thinking. For instance, a highly creative human response might include a diverse and unexpected collection of words such as "galaxy, fork, freedom, algae, harmonica, quantum, nostalgia, velvet, hurricane, photosynthesis." The disparate nature of these words signifies a high degree of divergent thinking, moving beyond conventional semantic links.
The performance on the DAT has been consistently shown to correlate strongly with results on other established creativity tests, including those used in creative writing, brainstorming, idea generation, and complex creative problem-solving scenarios. This suggests that while the task is language-based, it taps into broader cognitive processes that underpin creative thinking across a multitude of domains, making it an excellent proxy for general creative ability. Furthermore, the DAT offers significant practical advantages: it is quick to complete, typically taking only two to four minutes, and its online accessibility facilitates large-scale data collection from the general public, as demonstrated by the study’s impressive participant count.
Beyond Word Lists: Extending AI Creativity to Complex Writing Tasks
Recognizing that success on a word association task, while indicative, might not fully capture the essence of real-world creativity, the researchers extended their investigation. They sought to determine whether AI’s newfound proficiency in divergent thinking could translate into more complex and realistic creative activities. To this end, they pitted AI systems against human participants in traditional creative writing challenges, including the composition of haiku (a short, three-line poetic form renowned for its evocative imagery and strict syllable structure), the generation of compelling movie plot summaries, and the crafting of short stories.
The results of these more elaborate creative tests echoed the pattern observed in the DAT. While AI systems occasionally managed to exceed the performance of average humans, particularly in terms of linguistic fluency and the rapid generation of content, the most skilled human creators consistently delivered work that was not only stronger in narrative coherence and emotional depth but also demonstrably more original and truly imaginative. This suggests that while AI can master the mechanics and even the stylistic elements of creative writing, the spark of genius, the profound originality, and the nuanced understanding of human experience that defines top-tier creative output remain a uniquely human attribute.
Modulating AI Creativity: The Role of Temperature and Prompt Engineering
The study also delved into a critical question regarding AI’s creative potential: Is AI creativity a fixed attribute, or can it be deliberately shaped and adjusted? The findings revealed that creativity in AI is far from static; it can be significantly influenced by altering technical settings, most notably the model’s "temperature" parameter. This parameter acts as a control mechanism, dictating the predictability and adventurousness of the AI’s generated responses.
At lower temperature settings, AI models tend to produce safer, more conventional, and highly predictable outputs, often adhering closely to learned patterns and common associations. As the temperature is increased, however, the responses become progressively more varied, less predictable, and more exploratory. This allows the AI system to venture beyond familiar ideas and established norms, generating content that exhibits greater divergence and originality. This finding highlights a crucial aspect of human-AI interaction: the "creativity" observed in AI is not an intrinsic, unchangeable trait, but rather a tunable dimension responsive to human intervention.
Beyond technical parameters, the researchers also discovered that AI creativity is profoundly influenced by how instructions are formulated—a practice known as prompt engineering. For instance, prompts specifically designed to encourage models to delve into the origins and structural aspects of words using etymology led to more unexpected associations and consequently, higher creativity scores. These results underscore a fundamental principle: AI creativity is heavily dependent on human guidance and contextualization. This makes the art and science of interaction and prompting a central, indispensable part of the creative process when leveraging AI tools. It reframes AI not as an autonomous creative entity, but as a powerful amplifier whose output is intricately linked to the quality and intent of human input.
Broader Implications: Will AI Replace Human Creators?
The implications of this landmark study extend far beyond the scientific community, touching upon societal anxieties and industrial transformations. The findings offer a balanced and evidence-based perspective on the widespread fears that artificial intelligence could eventually replace creative professionals across various sectors. While the study unequivocally demonstrates that AI systems can now match or even exceed average human creativity on specific, well-defined tasks, it simultaneously highlights their clear limitations and their inherent reliance on human direction and calibration.
Professor Karim Jerbi addressed these concerns directly, advocating for a shift in perspective. "Even though AI can now reach human-level creativity on certain tests, we need to move beyond this misleading sense of competition," he asserted. "Generative AI has above all become an extremely powerful tool in the service of human creativity: it will not replace creators, but profoundly transform how they imagine, explore, and create—for those who choose to use it." This statement encapsulates the study’s overarching message: rather than heralding the end of creative careers, AI’s advancement suggests a future where it functions as a potent creative assistant. By rapidly generating diverse ideas, expanding conceptual possibilities, and opening new avenues for exploration, AI possesses the potential to significantly amplify human imagination and innovation, rather than render it obsolete.
The study’s findings resonate deeply with ongoing discussions in creative industries. Artists, writers, designers, musicians, and architects are increasingly exploring how AI can augment their processes, from generating preliminary concepts and drafting early iterations to exploring stylistic variations and optimizing creative workflows. The ability of AI to handle the more routine or computationally intensive aspects of creativity could free human creators to focus on higher-order conceptualization, emotional resonance, and truly groundbreaking originality—areas where their advantage remains undisputed.
From an educational standpoint, the study challenges traditional notions of teaching and assessing creativity. If AI can achieve average human creativity, then educational systems must adapt to foster the "elite" creativity that AI cannot yet replicate. This might involve emphasizing critical thinking, ethical considerations in AI use, and advanced prompt engineering skills, alongside traditional creative disciplines. Policymakers and intellectual property lawyers also face new challenges in defining originality and ownership in an era where AI can generate content indistinguishable from average human output, necessitating new frameworks for copyright and attribution.
"By directly confronting human and machine capabilities, studies like ours push us to rethink what we mean by creativity," concludes Professor Karim Jerbi. This reflection is crucial, as the boundaries between human and artificial intelligence continue to blur. The research does not merely provide answers but prompts deeper philosophical inquiries into the nature of consciousness, originality, and the unique contributions of the human mind in an increasingly technologically mediated world. It positions AI not as a competitor, but as a catalyst for a more profound understanding and perhaps even an evolution of human creative potential.
About the Study and Its Esteemed Collaborators
The seminal paper, aptly titled "Divergent creativity in humans and large language models," was officially published in Scientific Reports on January 21, 2026. This comprehensive research was the product of a collaborative effort involving leading scientists and institutions from across the academic and AI landscape, including the Université de Montréal, Concordia University, the University of Toronto Mississauga, Mila (the Quebec AI Institute), and Google DeepMind.
Professor Karim Jerbi of the Université de Montréal served as the principal investigator and driving force behind the study. He was supported by a dedicated team, with Antoine Bellemare-Pépin (Université de Montréal) and François Lespinasse (Concordia University) making significant contributions as co-first authors. The research team also notably included Yoshua Bengio, a figure of immense stature in the field of artificial intelligence. Bengio is the founder of Mila and LoiZééro, and is widely recognized as one of the pioneers of deep learning, the foundational technology underpinning modern AI systems like ChatGPT. His involvement underscores the profound scientific rigor and forward-thinking nature of this groundbreaking investigation into the frontiers of artificial and human creativity.
