Presented By
TIME 2030
Lon Tweeten for TIME; Getty images

Iyengar is a professor at Columbia Business School and the author of Think Bigger: How to Innovate.

Let’s face it. We’re more distracted than ever. Why remember anything when I can just Google it? Why summon the attention to read a book when I can just scroll through Twitter?

Some philosophers believe that ChatGPT and its siblings will further diminish our ability to do the kind of “deep work” needed to spark creativity and breed big ideas. What good are the tools if we begin to rely on them so much that we no longer have the capacity to think bigger? This argument is tempting because it’s romantic. If creativity is essentially human, there is something inherently limiting about the prospect of man replaced by machine. But the evidence tells a different story.

While seemingly “superhuman” technology can be intimidating, it generally enables us to become more creative — not less. In 1997, when the computer program Deep Blue beat the invincible grandmaster Gary Kasparov in chess, many feared that humans would begin to abandon the pursuit of chess mastery because they’d “never be as good as a computer.” In fact, the opposite happened. The widespread adoption of computer simulations made human chess players better. A recent study conducted by Henning Peinzuka of INSEAD found that in those countries where humans had access to computer chess simulations, their performance in chess improved. The players still found it useful to play against humans, but the presence of the non-human made the human a better, more creative player.

More from TIME

Now let us imagine the future of creativity in a world of generative AI that enables us to map choices as never before—to explore exponentially more combinations of choices, compare and contrast infinite approaches at a glance, and constantly test new ideas.

As the brilliant French mathematician Henri Poincaré once said: Invention consists in avoiding the constructing of useless contraptions and in constructing the useful combinations which are in infinite minority. To invent is to discern, to choose.

AI will not necessarily come up with our best ideas for us. But it will greatly reduce the cost—in time, money, and effort—of generating new ideas by instantaneously revealing untold options. It will enable us to efficiently discard the “useless contraptions” that cloud our vision and identify useful combinations previously unseen. It will empower us to broadly and efficiently canvas an incredibly vast range of domains to pull relevant knowledge from unexpected places. If used properly, AI will ultimately help us seed far greater innovation throughout our society.

Read More: How the World Must Respond to the AI Revolution

So how do we do it? How can we use large language models like ChatGPT to make us more creative? It starts with mapping our choices.

I often direct my students to perform a simple exercise that demonstrates the power of choice generation. I ask them to take two minutes to come up with as many answers as possible to a simple prompt like “Ways to use a toothpick.” After they share their ideas, I ask them to repeat the exercise; they almost always come up with more ideas during the second go-round than the first. Their creative juices have begun to flow. I ask them to do it again, and again. Inevitably, their rate of creation slows and the flow of ideas becomes a trickle. Their creative energy has become exhausted.

Now let us introduce ChatGPT. When I type “List the ways in which one could use a toothpick.” It instantaneously spits back 50 options. Here is but a brief selection:

Testing cake doneness: Inserting a toothpick into a cake to check if it’s fully baked.

Appetizer holder: Skewering small food items like olives, cheese, or fruit for easy serving.

Nail art: Applying small dots or lines of nail polish for intricate designs.

Cleaning small crevices: Reaching into tiny spaces to remove dirt or debris, such as in a keyboard or around jewelry.

Glue applicator: Spreading small amounts of glue for crafts or DIY projects.

Paint mixer: Stirring small amounts of paint for model-building or artwork.

Plant support: Providing extra support to small or fragile plants as they grow.

Type in “list more uses,” and the model spits back another 50 potential options. Some of the ideas are good, some not so good. The point is that the Chatbot can instantly find, collate, and list seemingly infinite possibilities that have already been created by humans across space, time, and context. Now students can apply their creativity toward assembling old ideas in new ways.

As Mark Twain wrote to Helen Keller, “substantially all ideas are second hand, consciously or unconsciously drawn from a million outside sources.” Innovators are most often strategic copiers who learn from examples of success, extract the parts that work well, and imagine new ways of using those pieces to create something new and meaningful. In the toothpick example, with ChatGPT the students no longer have to waste their time coming up with existing ideas; they can apply their creative energy toward iterating, assembling, and combining to create new, powerful ideas they would not have been able to generate without AI.

Now let’s take it a step further. If breakthrough ideas often come from unexpected places, how can we use ChatGPT to mine human knowledge’s vast hidden treasure troves to find the nuggets of knowledge that break our mental logjams? It’s easy to use the chatbot to map out choices within the same domain of query (i.e. If I’m looking to innovate on toothpicks, I use the chatbot to identify currently-known methods of using toothpicks so I can combine and iterate.)

But what if I start using the AI to map choices that are “out-of-domain,” i.e. from different times, different places, and across different industries? Suddenly our ability to think “outside the box” has increased dramatically. In fact, some of history’s greatest innovations come from inventors looking to entirely different domains to identify the various pieces needed to create something revolutionary.

Take ice cream, for example. In the 1840s, ice cream was only accessible to the very wealthy due to the high price of ice, the intense labor required, and the time it took to produce. Most of all, the freezer did not yet exist, so keeping the ice cream cold was enormously difficult. In 1843, a chemist and physicist named Nancy Johnson set out to bring ice cream to the masses by breaking the problem down, looking to history, and searching in new places for inspiration.

She started by searching for the ways other foods and beverages had their temperatures contained throughout history, which led her to pewter metal. By the Middle Ages, long before Johnson’s time, certain inns used pewter for mugs to keep beer and ale cold. She replaced the ceramic used to make ice cream at the time with cheap pewter and set it in a wooden bucket with a layer of ice packed around it to keep the mixture cold. Put on the pewter lid when you’re done, and your ice cream stays cold for hours.

Nancy still faced the challenge of stirring a mixture of cream, sugar, and other flavorings for hours on end. Was there a simpler and faster way to continuously mix the ingredients with less arm power? To remedy this, Johnson added a hand crank—an invention which went back to first-century China. From there, it spread to the Roman empire and on to the rest of Europe. The Eastern Mediterranean even implemented hand cranks to grind spices and coffee. In this application, the hand crank dramatically cut the time and effort it took to stir the ice cream in Johnson’s new contraption.

If we adapt Nancy’s approach to present-day problems, we can use ChatGPT to search out-of-domain in seconds. Say I’m an airline executive looking to improve customers’ experience at the airport. Sure, I could ask ChatGPT to spit out the various approaches airports have employed to improve the travel experience, but this list remains “in the box.” But what if I ask ChatGPT to list out examples of other experiences in which people are harried and upset. Here’s a brief selection: “Hospitals, traffic jams, courthouses, banks, the DMV, and funeral homes.” Now I can research tactics and precedents employed within each of those domains, pull out promising ideas, and combine and test to come up with a truly creative approach that might work for airports. From funeral homes, for example, I could draw on the power of empathy and comfortable environments and apply it to the airline gate experience. From hospitals, I could draw on methods for patient advocacy experiences and apply it to travelers. From the DMV, I could draw on attempts to bring more of the customer experience online and on mobile devices. Now I am working with a much richer and diverse set of elements to stir innovation.

These are but a few of the simple methods we must explore to harness the power of ChatGPT and its ilk to unleash creativity and widen our aperture to see a new horizon. The toothpick exercise is an example of infinite possibilities made new in real time. The ice cream example demonstrates the power of a historical lens to make the seemingly quixotic practical. And the airline example uses the chatbot to employ a powerful roving eye to inspect the “out-of-domain” world. As with any new technology, its power and consequences come down to how you use it. And the next time you need to “brainstorm” with ChatGPT, see what happens when you employ these methods; I think you’ll find you’re a lot more creative than you thought.

More Must-Reads From TIME

Contact us at

You May Also Like