My collaborators and I have been studying “creativity” in non-human animals, such as dolphins, killer whales, and dogs, using a training technique that promotes variable non-verbal behaviors. (Watch this short video to learn about our marine mammal research.) Using this training technique, dubbed “innovate” or “create”, trainers have taught a variety of animals to produce a different behavior each time the animal is given a specific signal (a word or hand gesture). Sometimes, the animals produce completely novel behaviors, which can be increased through the use of well-timed reinforcement.
Karen Pryor recounted her experience training two rough-toothed dolphins (pictured below) on this “creative” activity in her book “Lads Before the Wind: Diary of a Dolphin Trainer” and its initial scientific paper. As Pryor wrote, the “innovate” behavior did not use shaping to create novel behaviors but rather provided feedback for the animals on when they produced a behavior that was initially different, and eventually, new to them.
Creativity in humans is usually measured artistically or verbally with standardized prompts given within a certain period of time. These “tests” provide opportunities to evaluate the produced responses for different characteristics, such as how many different responses can be produced (fluency), how many different categories responses can represent uniquely (flexibility), how many different details were added (elaboration), and how novel or distinct the responses are (originality).
Since animals can’t draw, speak, or write, behaviors such as “innovate,” spontaneous play behavior (see linked video of the same dolphins from Anthony’s Key, Roatan, Honduras 10 yrs before creativity video), or foraging methods (see linked video of “conching” bottlenose dolphins in Shark Bay Australia) have been offered as ways to evaluate creativity in nonverbal individuals (including young children). These studies and various perspective articles have provided examples of animals (dogs, bottlenose dolphins, killer whales) producing original actions under specific conditions, such as Pryor’s innovative paradigm using reinforcement and inventing novel foraging solutions, like killer whales working together to access seals on icebergs.
In a recent article by researchers Dr. Daniel Oppenheimer and Dr. Mark Patterson (pictured below) from Carnegie Mellon University, a new avenue of creativity was explored with university students. While animals have learned how to use computers, they haven’t quite figured out how to use the internet to search for topics of interest. In a study published by the Psychonomic Society’s Memory & Cognition and summarized in this post, the researchers sought to determine whether using the internet facilitated or inhibited creativity in university students.
Previous research had produced mixed results for generating creative ideas. For example, working in groups and pre-exposing individuals to ideas can constrain the number of ideas generated; yet, some studies suggest that utilizing additional online resources can stimulate idea generation, potentially through priming or by offering rare options. Prior to the study summarized in this post, almost 1,000 publications investigated the relationship between digital tools and creativity, with the majority evaluating this question at the individual level.
Oppenheimer and Patterson chose to investigate their question using the alternative uses task within the context of two objects, an umbrella and a shield. This task requires participants to come up with as many uses as they can for a particular object, such as a block. In the current study, the participants were instructed to conduct a Google search or not within a 3-minute period for one of the two items given. All participants were randomly assigned to one of four conditions: Google search-umbrella, Google search-shield, no Google search-umbrella, and no Google search-shield, resulting in a fairly even distribution across the four groups.
The results at the individual level showed no difference between the Google-search and the No-Google search groups for “shield” uses. However, being able to conduct a Google search significantly increased the number of uses generated for “umbrella”. This result emphasizes the fluency variable of creativity.
In contrast, when groups were simulated using different sample sizes, some differences emerged. (I considered this set of analyses a creative application of statistical simulation methods.) Referred to as a “nominal group,” since the groups are randomly created and do not represent a true collaborative group working together in real time, the No-Google-search group (turquoise line) created more distinct uses for a shield than the Google search group (red line), with no limitations on sample size, as seen in the top panel in the figure below.
However, when the umbrella term was used, the Google-search group (red line) outperformed the No-Google search group (turquoise line) for sample sizes of less than nine individuals, as shown in the lower panel in the figure above. As the authors concluded, Google searches that produce many uses can be helpful when “nominal” groups are small, but are limited when internet use is limited or groups are large in size.
Distinctiveness was also assessed for the responses provided. Once again, individuals with internet access were more likely to report common uses or less distinctive uses for the familiar object—the umbrella. The less familiar object, the shield, produced more distinctive uses, whether with or without googling.
These findings were upheld when additional subjective coding was performed on the individual responses by two independent raters who evaluated all responses on a 5-pt scale for three different dimensions: (1) effectiveness or how useful the purpose created by a participant was, (2) novelty or how new or surprising the created purpose was, and (3) “creativity” or an overall rating of creativity. The figure below shows the results of the two coders (circles & triangles, respectively) for the three dimensions they coded. As can be seen in the figure, the coding for each of the dimensions tracked the simulation results for both objects, despite coder 2 (triangles) being a more stringent coder (fewer 5 ratings) than coder 1 (circles).
Ultimately, the researchers found that the qualitative assessments generally validated the findings from the simulations performed with the nominal groups, but with the caveat that groups formed nominally using Google (red lines in figure above) were less creative, less novel, and had fewer effective solutions than groups not using Google (turquoise lines in figure above).
Using this creative approach (i.e., creating “artificial” groups of different sample sizes statistically), Oppenheimer and Patterson reevaluated data from a similar previous study in which no difference in output was found between individuals who used internet searching versus not for four different objects: brick, nail, paperclip, and safety pin.
Utilizing the same coding and statistical methods from their study with “umbrella” and “shield”, Oppenheimer and Patterson’s re-evaluation of the secondary data found evidence that as “group size increased, a creative advantage for non-Google nominal groups emerged.” Oppenheimer and Patterson further concluded that the reason for the null findings of the previous study was most likely related to a methodological constraint: participants could only provide up to 5 uses, rather than an unlimited number of uses, within a specific time frame (i.e., 3 minutes in the study summarized in this post).
As Oppenheimer stated,
“The internet can expand creativity by surfacing ideas that people might otherwise never have known about or thought about, but being exposed to ideas can constrain creativity because we tend to think in ways that align with ideas that we have been primed with. If everybody uses the same sorts of digital technologies, everybody will get primed to think in the same ways, and that can constrain the space of ideas that groups of people come up with. The very same features of the internet that can boost creativity in individuals can reduce creativity for collectives, and that has important implications as we try to generate solutions to the big problems society faces.”
As the American Romantic poet James Russell Lowell stated in the 19th century, “Creativity is not the finding of a thing, but the making something out of it after it is found.”
I suspect Oppenheimer’s words of warning will remain especially critical as AI becomes more pervasive.
Featured Psychonomic Society Article
Oppenheimer, D. M., & Patterson, M. T. (2025). Thinking outside the box means thinking outside the search engine. Memory & Cognition, 1-14. https://doi.org/10.3758/s13421-025-01732-x
