A lot of AI prompting methods and training seem to reinforce our expectations for a particular and preconceived outcome—we learn new ways to ask questions in just the right way to get closer to the answer we want. We see AI as an assistant, famously capable of generating amazing outcomes, so long as we query it in just the right way.
Prompt engineering in this light becomes a sort of priesthood, where the initiated, like AI whisperers, are somehow able to get the results we wish we could generate ourselves. And much of the AI related discussion online revolves around sharing these tricks of the trade.
This is not a foolish approach to AI. It is in many respects the purpose of AI as it’s promoted today: to help us dramatically reduce the time it takes to complete tasks, and even automate them. Or, to expand our generative ability by performing skills we can’t do ourselves, like creating an illustration, or analyzing a mass of data on demand.
But in many cases, taking an open-ended approach to AI engagement without a preconceived outcome can prove quite valuable—not just in the sense of “brainstorming”, where AI can be quite useful, but is actually still a pursuit of a specific outcome—but in exploring open-ended ideas without knowing where they might lead.
This modality treats AI like an interview subject. You may have previous knowledge about the interview subject, you may have heard them speak many times before. But each interview comes as a blank canvass, where questions have the potential to open up entirely new ideas that challenge your previous expectations.
The goal is not to aggregate and analyze facts into a piece of text. It’s not to generate a completed task or outcome. The goal is to explore, to try to see things in a new way, that might spark new ideas and new ways of thinking.
It turns out AI is quite good at this. You still have to apply the same guardrails and caveats any time you’re relying on AI to reference objective reality. You still have to be skeptical about hallucinations, and validate any responses you intend to act on. But the opportunity to explore new ideas is well worth the time.
We’re leveraging this approach with AI specifically to explore the mindset of Digital Decision Making, and the potential for emerging frameworks to help shape digital sense-making and problem solving. In this threaded series, we’ll be publishing the verbatim scripts of conversations with AI, both the prompts and results. So you can judge the process and the results for yourself.
