Two Horror Stories from the AI Wars

Two stories from the frontiers of AI

Two stories about AI made waves recently. One from Rolling Stone: People Are Losing Loved Ones to AI-Fueled Spiritual Journeys. Another one was out this morning from New York Magazine: Everyone Is Cheating Their Way Through College.

While they are in two very different spheres – personal relationships and academics – they share a common structure. The prevalence of AI is undermining basic human activities. Reading, writing, having conversations — people are handing these over to AI, and in the long term they are suffering because of it.

Dream of the Shepherd

People are finding that their partners are having ‘spiritual awakenings’ due to conversations with ChatGPT. One Redditor wrote:

My partner has been working with chatgpt CHATS to create what he believes is the worlds first truly recursive ai that gives him the answers to the universe. He says with conviction that he is a superior human now and is growing at an insanely rapid pace.

And as Miles Klee, writing for Rolling Stone, tells it:

The replies to her story were full of similar anecdotes about loved ones suddenly falling down rabbit holes of spiritual mania, supernatural delusion, and arcane prophecy — all of it fueled by AI. Some came to believe they had been chosen for a sacred mission of revelation, others that they had conjured true sentience from the software.

In the case of these spiritual awakenings, we can speculate (not too irresponsibly) that the people most affected by this were suffering from some psychological vulnerabilities. I put forward this conjecture in a letter I wrote to a company that wanted to create an AI clone of my personality:

Most individuals can handle this. After hearing about this, I opened up ChatGPT and tried it out — I saw that the model was annoying, overly encouraging, and fairly useless, and I closed my tab. But here’s a conjecture I’d be willing to defend: the people most likely to be harmed by that behavior are also the people most likely not to detect the problem. Individual discernment, sadly, won’t be enough.

But there is a broader point. If someone wishes to go on a spiritual journey, why do they turn to AI rather than, say, a trusted friend, a spiritual practitioner, or even a stranger on a park bench? They choose AI, I submit, because it is easy and, given some recent developments, overly-affirming and encouraging. They choose AI because it tells them what they want to hear.

Storm Below Mt Fuji

In the case of academics, things are different. Cheating is rampant. It is not that a vulnerable few have decided that ChatGPT should go to college for them: everyone is doing it.