The extremes of AI

At this moment, AI feels like a battle in the extremes, a war where you HAVE to choose sides. You're either team #truebeliever, believing that everything in the world must be solved or approached using AI, or you're team #resistAI, believing that AI is the downfall of humanity and should be annihilated.

It feels like there's no nuance in this debate, no room for someone like me who sees AI as a tool, a step forward in the long road of human technological innovation, but not some earth-shattering, humanity-ending, job-decimating discovery.

Much of the AI hype feels companies trying to bolster their position in the market, less about “AI is changing the world” and more “AI is gonna make us rich if we hype it!” When companies go public with a nearly $1 TRILLION valuation, it's a hard reminder what this is really all about. I mean, Sam Altman hasn't exactly turned out to be the altruist he wanted people to think he was. And don't forget the YouTubers, data center builders, and every other AI-adjacent person/company that is also on this gravy train of AI-generated slop.

I work in a knowledge-worked type of job, and we're struggling with what AI means for our profession - what it will replace, what it won't, what efficiencies it can bring, what it will cost (actually and metaphorically). During recent discussions with my colleagues about this, quite a few people had clearly bought into the AI dichotomies of the extreme. In each conversation, people had either sworn off AI entirely, refusing to even look at it lest it consume their soul, or gone down an AI rabbit hole so deep they were almost resigned to being replaced by a computer terminal in the very near future.

But a handful of people articulated what I had been thinking:

Why does it have to be one or the other? Why can't it be somewhere in the middle?

That’s where I am. I am the (not-so) silent middle.

I guess this is the problem with debates; you're forced to pick a side. Luddite or zealot. A dinosaur or an acolyte. There's no room for nuance in these positions, no room for rational discussion.

I'm not afraid of AI, I don't think it will consume my job and leave me penniless. I don't think it will annihilate the human race. The internet fundamentally changed our world, as did internet search engines and the iPhone. I lived through all of these changes. In my childhood, a computer was a behemoth that sat in the living room, took many minutes to boot up, and required DOS prompts to start a game. Now, my computer is a much more powerful slab of glass that fits in my pocket. This was a fundamental change, a shift in how we lived and navigated the world as humans.

But humans kept marching forward, adapted. For any job or profession that was lost, new ones were created. I mean, who the hell thought "Instagram Influencer" would ever be a profession?

That said, we as a society need to have an honest conversation about what AI can and cannot replace. And more important, what is lost when we do replace something with AI.

A good example: one thing we do at work is draw conclusions based on evidence. We think critically about the evidence, about the conclusions, and about the larger context in which we want to say something. Let's say that AI eventually could replace this. What would be lost? I'd argue that the loss would not be worth the efficiency, because we'd be losing the human nuance in the process that I don't think AI could ever replace.

I recently had AI help perform a task at work, a task that was rather mundane, but kind of informative. It was a task that took time, a task that in the past I would have delegated to an entry-level staff person. The problem is that budget cuts where I work means we no longer have entry-level staff. Yes, AI can perform the task reasonably well, but there is a loss here - a young person is not being trained to eventually replace me, a young person is not getting the knowledge that I have acquired over many years performing similar tasks. Because mundane tasks that can be outsourced to AI are the foundation on which higher-level skills and bigger ideas are built

AI isn't going to "transform" medicine, but it will bring about efficiencies and improvements. That is clear. But what is lost if we tried to replace too much of our health care with AI? AI can consume facts about a patients - descriptions of symptoms, lab values, radiology imaging. But AI can't visually see the patient and note an unkempt or neat appearance that might signal to a health care provider level of functioning. AI can't tell what level of support a patient has in their life, which may be apparent based on whether the patient attends appointments alone or with a loved one. AI can't discern subtle cues from a patient about what is and is not important to them as a person, desires that may change the decision-making process for an illness.

We have a lot to think about and discuss in the great AI debate, but to do that we have to stop approaching this like opposing teams and instead act like the intelligent humans we’ve purportedly evolved to be.