AI Paralysis

I've been talking with a lot of design leaders lately, and there's a pattern I'm seeing (or should I say a paralysis) that I find worrying related to AI. In a nutshell, a bunch of really capable teams are stuck right now over how and where to integrate AI into their workflow, and not because of the technology itself (though that too can be overwhelming) but because of the pressure from upon high to be seen doing something with it as some sort of savior to their respective companies. They can feel their boss or their CEO pressuring them about AI use but they're in limbo, because though they know they're "supposed to be doing something with AI," nobody's aligned on where to aim the effort.

I feel like I've seen this movie before. Remember back when every brand on earth wanted a viral social post (that still happens, BTW)?!  Not because going viral did a single thing for their kind of business, but because it's what the cool kids were doing and no exec wanted to be the one who wasn't. A lot of money and wasted time got poured into chasing a thing few ever bothered to tie to an actual outcome.

AI feels sort of that same enticement, except the consequences are higher and the FOMO is way worse. Because with that kind of pressure, there is no time to slow down and ask what we are trying to use these shiny new tools to solve, improve, speed up, etc. Even just a small pause feels like falling behind, but without taking a minute to ask where the effort is best focused, it becomes way too easy to fall down unnecessary rabbit holes creating flashy visuals or automating mundane tasks instead of actually figuring out where to make a real impact.

To be clear, this is not me suggesting we avoid AI, quite the contrary. I'm genuinely excited about AI and use it for a number of tasks pretty much daily. This is me saying we need to point it at real problems first. The teams I've seen get the most out of all this weren't the ones noodling around (though that’s fun too), they’re the ones who stopped and asked "okay, what's actually broken here/what can be faster/what might this unlock, and how can an AI tool help with that?".

It’s funny, this often also feels a lot like CAD adoption in the 90s. Just like then, these are tools that are both exciting and terrifying, and yet the implementation discipline of the new technology is kind of the same today as it's always been.

I think it’s safe to say, by now we’ve all seen AI output that’s blown our minds just as much as we’ve seen useless $#it. My challenge for us is to ask what we're hoping to solve or improve with AI tools before we fall in love with what comes out the other end. 🤔