“Well, at least the robots won’t train horses,” I joked to a fellow English teacher as we chatted about the rise of Artificial Intelligence-generated college essays and programs like Moonbeam or ChatGBT.

“Why not?” she replied.

I quipped back some platitude about “feel,” but honestly, this question kept me up at night for months. Yes, as the rest of my colleagues lost their minds about job security and the future of American Letters, I was freaking out about ponies.

If I’m a bit blasé about the “threat” of AI, it’s because I’ve heard the “Wolf!” cry one too many times. The “death” of the humanities, especially its literary aspect, has been repeatedly declared throughout millennia. In my lifetime alone, the internet, text messages, and e-books have all been called the doom of writing as an art form. While business schools get another new wing, the building where the English departments are housed seem to look a little sadder with each passing semester.

Yet somehow, books are still published, people still read, and some even still write. For all we know, AI-generated writing could pass with a blip like Google Glass or move at the glacial pace of progress like the driverless car.

As someone with dyslexia, I will argue that AI can spell and proofread far better than I could ever dream of. As someone with multiple English degrees, I’ll counter that what it does poorly is say anything particularly interesting or challenging; its aim is to people please, to barf out generic answers, often regardless of accuracy or nuance.

So for those who care most about immaculate mechanics, I am glad you have found ChatGBT. For everyone else, humans will be here to write you something more engaging and thought provoking.

I’m less nonchalant about AI when it comes to my hobby, however. While I may have been prepared for AI to come for my job, I thought my horses were untouchable.

Now, I’m not so sure.

Horses thrive on patterns and predictable behavior. If they get the exact same cue for something and then the exact same feedback for it every time, they learn faster. “They are consistent” is among the highest praise I can give a fellow equestrian. And consistency is the forte of Artificial Intelligence.

AI doesn’t feel fear. It doesn’t need health insurance. It doesn’t get distracted by its life out of the saddle. Maybe it can even be programmed so that it doesn’t develop my bad habit of grabbing the inside rein when I get freaked out. Perhaps it is the rider our equine companions deserve?

The whole thing makes me nauseous.

Yet perhaps I am asking the wrong questions. What if the question of AI isn’t whether or not horses will need us less, but rather will we need them more?

In a recent episode of his podcast, The Ezra Klein Show, Klein opines that humanity may soon “find ourselves taking metaphysical shelter in the subjective experience of consciousness, the qualities we share with animals.” In a world of almost godlike AI, in other words, our status as feeling, subjective, flawed, and vulnerable meat sacks is the thing that will set us apart. AI will probably never be an animal, but we will always be, and horses are adept at reminding us of that fact.

Unlike AI, horses won’t always spit back the information we want to hear. They often tell us things about ourselves we would rather not think about. Horses, unlike AI, are reliant on their senses, their emotions, and their relationships—they are the constant reflection of our own fleshed and flawed existence.

I was reminded of this during a recent lesson when my horse spooked at the snow sliding off the roof of the indoor arena. As one of the larger chunks of snow thundered down the tin roof, my mare did not one but two 360s. All of a sudden, my dressage lesson turned into a reining one.

The adrenaline and the overwhelming sense of pride that I stayed on momentarily stunned my seasonal depression into submission. I grinned, terror and joy pulsing in my ears. After a while, I lost track of how many times she spooked and how many times I sat back and told my nerves to sit in the corner and think about their behavior.

For that horse and me, our connection deepened after that ride. We “survived” that harrowing moment together and rode away more confident that we could get through more such moments in the future. Our relationship deepened further because of our mortality.

I don’t think a connection that profound and hard-won could ever be generated by a bot—even if it is better at spelling than me.