Most of work life doesn’t involve that much creativity. Instead, it involves experience. The more that you have seen before, the more likely it is that you will have a ready solution for something that comes up at work, because you have previous experience to draw on. That is effective, but it isn’t really that creative.
Creative problem solving is needed when the situation doesn’t remind you of anything you have seen before that will provide you with a course of action. Of course, your first approach to attack this question should be to ask for help from someone with more experience than you. They may know something you don’t to address the problem.
If you and the people around you are all stuck, then you have to get creative. As I wrote in Smart Thinking, an important route to creativity is finding an analogy to the current problem that can provide you with insight about how to address it. That is where AI can help.
Distill the problem
Pick your favorite Large Language Model to help. (I have a Claude project set up for just this purpose.) First, have the LLM help you find the essence of the problem. The reason why analogies are hard to come up with is that the way you’re thinking about the problem is filled with specific details about the objects, people, and other elements that have caused the problem. That leads you to continue to be reminded of other things you know that involve similar objects, people, and elements.
To escape the tyranny of these specifics, you need to develop a more abstract description of the problem that captures its essence, but doesn’t refer to the specifics. For example, if you were trying to come up with a novel weed killer, you might focus on the essence of weed killing, which is one of selective destruction (because you’re trying to eliminate only the weeds and not the other plants around them).
Describe the problem to the LLM and ask it to help you find a more abstract description of it. Rather than having the LLM do all the work, have it ask you questions to help you find a more abstract description, and have it evaluate your attempts. Over time, you’ll find that you have developed a description you like.
Come up with analogies
From there, see what it reminds you of. The things you think of may be potential analogies. Returning to our example of a weed killer, you might be reminded of oncology drugs (which have to kill cancer cells, but not surrounding healthy tissue). When you are reminded of something, engage in a discussion with the LLM about whether that analogy provides any insights into the problem you didn’t have before. The LLM may also look up articles related to the domain of the analogy that you can read to learn more.
If you haven’t yet found a solution to the problem, you can then ask the LLM for any analogies it can think of. For any suggestions the LLM makes, you can go through a similar process of exploring the new domain to see how it makes you think differently about your problem.
AI as a creativity partner
This approach to using AI has two benefits. First, it gives you scaffolding for finding and using analogies, which is a powerful tool for creative problem solving. Second, by having the model support your thinking, you are also practicing your own creative skill. The more that you engage with an LLM as a support for your own effective thinking, the more that you will also be effective at finding good analogies without the tool.
The more general strategy here is that we often use LLMs to solve problems for us. The better approach is to use LLMs to support our own development of more sophisticated thinking techniques. This way of using AI has two advantages. First, it enables you to improve your skills. Second, the quality of the feedback an AI model can give you is related to the sophistication of the prompts you feed it. As the quality of your thinking goes up, you’ll get more nuanced engagement from the AI model.