Perfectly Written, Completely Wrong: AI and the Complexity of Being Human
As we're all grappling with the impact of AI on our futures (cue scary headlines about the fall of civilization), we're also exponentially exposed to the impact of AI on a daily basis. I mean, who hasn't at this point used AI to help draft an email or social media post? I've become suspicious of my own use of late, so I decided to take a short poll of friends and family to hear how they've been using AI in their daily lives, and here are some use cases I've learned about: a wedding toast, an angry customer review response, a review of a book, and even a breakup text! (The unnamed person who shared the breakup story reports his ex-girlfriend seemed to appreciate his articulation.) I get it — as I just stated, I'm not immune to the euphoria of this alien life form appearing out of nowhere to literally save us countless hours.
Personally, something's been bugging me about this human experiment we all find ourselves in, and maybe I haven't chosen to address it yet because it's all happening so fast that I'm missing the signs altogether. Or maybe I haven't because I don't want to admit that I'm willing to give away a little bit of creative expression and personal agency for the abundance of time I'm unlocking by accepting the inevitable presence of AI help in my daily life.
The catalyst that startled me into this particular self-dialogue about AI was actually a promising new professional relationship with a 21-year-old (whom I’ll call Evan to protect his identity). I had recently hired Evan to help manage social media for one of my restaurants. He was young and inexperienced, but he had passion and creative expression, and I liked him. He reminded me of me 20 years ago. I spent quite a bit of time with him in person setting expectations, talking through what I was looking for, and trying to give him an opportunity to express himself on his terms without getting in the way. I thought he had potential, and in working with him for over one month, nearly all of our communication had been face-to-face and positive.
Then one day, I received a text message from him that was so bizarrely different from every interaction we'd had that it immediately caught me off guard. In a surprisingly formal and aggressive tone, Evan informed me that he would not continue providing the services we'd previously discussed until I paid him what he was owed, that he wasn't interested in discussing the matter further until payment was made, and then proceeded to lay out several expectations for how things would work going forward.
The outstanding balance was $17!
As I read the message again, it became almost comical. This sounded less like the 21-year-old kid I'd been working with and more like a communication from an established social media conglomerate preparing to enforce the terms of a million-dollar contract. The text was perfectly polished, using punctuation I hadn't seen since my graduate-level days in college. And then it hit me: there was almost no doubt in my mind that I was reading something generated by AI.
What struck me wasn't that the requests (borderline demands) themselves were necessarily ridiculous. Given a vastly different set of circumstances, some version of them might have been perfectly reasonable. My immediate suspicion was that Evan had given AI a couple of facts about the situation and received back a confident, professional-sounding response given what little context his AI could have known. What the AI couldn't possibly account for, unless Evan took the time to explain it, was EVERYTHING else that I'd categorize as inherently human: the nature of our relationship, the opportunity he'd been given, our previous conversations, the relatively informal nature of our arrangement, the fact that we had almost exclusively communicated face-to-face, and, perhaps most absurdly, that the financial dispute at the center of this newfound hardline negotiating position was seventeen dollars.
And maybe that's what startled me most about the whole interaction. Evan knew all of those things. AI presumably didn't. Yet somehow, in that moment, the response generated from the thinner understanding of the situation won out. AI can produce something polished, confident, and remarkably convincing from very little information. But the polish of the response doesn't mean the judgment behind it fits the circumstances. We still have to recognize everything the machine doesn't know.
I don't have the answers to how we integrate with AI — I'm not sure I've heard a compelling argument yet on this account. What I can do is something I've always done — pay attention to the impact of the environment on me. Stay curious, stay grateful, and never take the opportunities that life offers too seriously, or I might miss out on the human experience.
For those of you wondering, I called Evan after receiving the text. I've yet to hear back from him, but I'm hopeful we'll get the opportunity to actually speak like it's 1999 at some point.
