Here’s how to use AI to fuel creativity instead of destroy it

A few years ago, I discovered a tomato sauce recipe that was surprisingly simple: just canned tomatoes, butter, salt, and an onion. It inspired me to experiment, adding this and that each time to see how the flavor changed. Today, I’d call myself an amateur sauce expert. I know exactly how long it needs to simmer, what shade of red signals it’s ready, and how to improvise with whatever’s in the fridge.

As my kitchen exploits remind me, experimentation is part of learning. It wouldn’t be the same if I’d just asked ChatGPT how to make sauce each time. I’d be outsourcing my culinary creativity and losing the teachable moments that come from trial and error.

As New Yorker writer Joshua Rothman observed, “[I]t’s becoming clear that artificial intelligence can relieve us of the burden of trying and trying again. A.I. systems make it trivially easy to take an existing thing and ask for a new iteration.”

AI can boost creative thinking—or eliminate it entirely. As the CEO of a company built on automation, I’ve found that the key is to treat AI as a creative collaborator, not a replacement.

Here are a few rules of thumb for striking the right balance.

1. Use AI for idea generation—not final decisions

When generative AI became widely available, a lot of hype swirled around its implications. Professionals, from knowledge workers to authors and beyond, feared that AI would take their jobs. AI seemed destined to keep improving, outpacing the skills and intelligence of its human counterparts.

More recently, the technology’s limits have become more apparent. While AI tools remain powerful workplace tools, their progress is unlikely to be endlessly exponential. As Cal Newport notes, critics argue that “the technology is important but not poised to radically transform our lives;” it may not get dramatically better than it is today.

AI won’t write the next great novel or compose symphonies to rival Bach. But it is an excellent brainstorming partner. Wharton professor Christian Terwiesch, who tested ChatGPT’s idea generation against college students’, explained: “It’s cheap. It’s fast. It’s good. What’s not to be liked? Worst case is you reject all of the ideas and run with your own. But our research speaks strongly to the fact that your idea pool will get better.”

Let AI tools like ChatGPT help you generate more, better ideas. Start with your own thoughts, and use AI to generate alternatives. Then, apply your own human judgment to refine and select the best path forward.

Treat prompts like a conversation, not a command

One of the strengths of generative AI tools like ChatGPT is their conversational nature. Think of your first prompt as an icebreaker—it’s just there to get the dialogue started.

While I recommend being as specific as possible, that is, giving the tool enough context to generate strong, accurate replies, you can always refine as you go. To offer a visual, imagine your dialogue with AI as a funnel: wide at the top and narrowing as you move toward the bottom.

You might start by asking ChatGPT to generate ideas for a marketing campaign. Once it produces a list, ask it to refine those ideas for a specific target audience—say, tech entrepreneurs in their 20s to 40s, or suburban parents.

Keep iterating until you land on the output that works best for you.

Create space for experimentation

Curiosity may have killed the cat, but it certainly won’t kill your company. In fact, a healthy sense of curiosity among employees will strengthen it. Research-backed benefits include boosted innovation, reduced group conflict, fewer decision-making errors, and improved communication.

While leaders often claim to value curiosity, they tend to stifle it, preferring that employees stay within the lines. Instead, leaders should give employees the freedom to explore. Build enough slack into their schedules so they can test and tinker with AI tools without the pressure to prove immediate ROI. Lead by example: Share your own experiments—whether it’s trying a new AI feature or recounting an automation gone awry. Getting it wrong can be valuable, too.

Ask for employee feedback to uncover what they’re curious about: new systems or tools they’d like to try, or better ways the company could operate. It might feel inefficient, like a misuse of employee time and effort, but in the AI era, staying competitive depends on curiosity and experimentation. Like a chef giving their team full access to the kitchen, leaders must create workplaces where both creativity and experimentation can thrive.