We’re Skipping the Thinking For the Cheap Thoughts

“What I cannot create, I do not understand.”

–Richard Feynman, CIT Professor & Nobel Prize winner

Let’s get this out of the way I use AI. My colleagues use AI. My clients use AI. I’m also not going to stop using AI. It’s useful. But it also really bothers me.

It’s taken a minute to figure out why. Sure, I fret over the future employment of my friends and family. I panic over Skynet. All the usual stuff, but AI has also always felt… I don’t know… icky.

We had the opportunity to help Common Sense Media relaunch their Digital Literacy and Wellbeing Curriculum, which had just been update to help teachers enable their students to think critically about AI. (It’s a really great organization. I’d recommend checking them out). In interviewing teachers one insight felt clearest: while knowledge has never been easier to come by, critical thinking has never been harder to instill; that unless we did something soon, we’d give up the ever important friction of learning.

That’s when the “a-ha” hit. I use quotes because this shouldn’t be a major revelation, but: this (our work) is supposed to be a long, arduous process. And AI is robbing us of that with the promise of ease.

I know you lose credibility when you start with “when I was your age,” but when I was your age we traveled 1000s of miles to take in how people lived in their homes, we spent days organizing Post-Its on foamcore to analyze what we’d heard, we’d spend late night agonizing over choice of words, and then, once we finished our work, threw it out and started over when our colleagues found the flaw in it. It was really hard. But the work was really good.

Just like students we need the friction of learning. We need to sit with heavy human questions and think about what they actually mean, and then sit and think about them some more. We need the moment our brain breaks and we’re forced to put pens down, only to be struck by a moment of brilliance on the walk home. When we force ourselves to do it the hard way, we get better results. The long, drawnout, painstaking process mattered, but with AI we risk losing this by skipping the thinking to get straight to the thought.

When the researchers compared whether process or analysis was more important in producing good decisions—those that increased revenues, profits, and market share—they found that “process mattered more than analysis—by a factor of six.” Often a good process led to better analysis—for instance, by ferreting out faulty logic. But the reverse was not true: “Superb analysis is useless unless the decision process gives it a fair hearing.”

–From Decisive by Chip Heath & Dan Heath

Here’s the thing, though. I’m not going to stop using AI. And you probably shouldn’t either. But I implore all of us in the knowledge field to not give up on process. Find efficiency but don’t lose the difficulty. Because if we do I’m almost positive Claude thinks Crystal Clear Pepsi is a winning idea.

Anywho, thanks for reading this far. I hope you have a long and arduous work week.