AI Now Beats Average Humans at Creativity Tests. The Top 10% Still Win.

A study from the Université de Montréal tested over 100,000 humans against today's leading AI models using standardized creativity tests. The result is striking: generative AI outperforms the average person, even using older models like GPT-4.

The study used the Divergent Association Task, a validated method for measuring creative thinking where participants generate the ten most semantically distant words they can think of. AI models consistently outscored 72 percent of human participants.

There is one critical caveat. The top 10 percent of humans still win. Human peak creativity, driven by lived experience, cultural context, and genuine curiosity, remains beyond AI's reach for now.

The implication for the workplace is clear. Jobs requiring average creativity, such as writing ad variations, product descriptions, or headline alternatives, are already in AI territory. Jobs demanding unusual, deep, or personal creativity remain safe for the foreseeable future.

For companies like Wenaas and similar brands, this means baseline product copy and marketing variations can be automated effectively, while storytelling and brand voice work with genuine human depth should still be done by people.