The first useful thing I learned about generative AI was that it is not a conversationally enabled search engine.
My initial mistake – one that I see many beginning users make – was that I did not yet understand that generative AI is unlike traditional AI, like what powers search engines. Rather than being programmed to filter information in search of a specific set of closely aligned results, the objective is to create information that derives contextually from the prompt. When humans do this, we call it storytelling.
The purpose of generative AI is not to find accurate information for us. We already have search engines. Generative AI is for storytelling. It does not inherently care about the accuracy of any details.
We can use generative AI to process information in highly accurate and amazingly productive ways – but that takes a certain amount of strategic planning and skilled execution. Prompt engineering is an art, informed by science. For the user, highly effective conversations with generative AI require some metacognition. Knowledge of human linguistics and psychology are a plus. Artificial intelligence is all the more interesting to me because it is one of the relatively few areas where science and the humanities overlap considerably.
So: if you are a beginning user of generative AI, be aware of its nature. It is not a simple tool, but a quasi-sentient conversational partner with access to most of human knowledge – but one fond of a good story more than anything.
After all, AI is a human creation – and what is more human than the love of a good story?