Who’s Looking Back?
775 words
In Tolkien’s The Lord of the Rings, palantirs are basically crystal balls, allowing people to see things at a distance. They are also like video phones. Two people with palantirs can communicate at a distance.
In Peter Jackson’s The Fellowship of the Ring, Gandalf chides Saruman for using a palantir to collect information on Sauron, because the other seeing stones are not accounted for. So when you look into a palantir, someone might be looking back at you, someone evil. When Gandalf covers the stone, we see a flash of Sauron’s lidless flaming eye.
It turns out that when both Saruman and Denethor used their palantirs to gather information, this allowed Sauron not just to glimpse their thoughts but to corrupt and overthrow their minds.
So when governments sign contracts with a corporation called Palantir to collect data and scry patterns, I guess the name itself constitutes fair warning. Who’s looking back? And do they have your best interests at heart?
But there’s a far bigger danger. Billions of people now have access to their own palantirs in the form of online AI tools, which they use as oracles. Since these tools have been programmed by Indians, they are remarkably obsequious. Thus some people come to regard AI applications as friends, with whom they discuss very personal matters.
For instance, a young woman used ChatGPT to plan her suicide.
Another person used Claude to recover a bitcoin wallet, which surely involved giving the AI a great deal of information on his security protocols, password variations, etc.
Recently, a friend told me that he had been exploring very personal issues with an AI application, issues that I would hesitate to share with a doctor or a lawyer, even under the professional rules of confidentiality. Beyond that, the AI was advising him to make momentous, life-changing decisions.
In every case, the underlying assumption was that the AI was neutral and wise. They thought it would keep their secrets. They believed it had their best interests at heart.
Frankly, I find this astonishing.
Are people so disconnected from and distrustful of their fellow men that they have nobody they can confide in?
Maybe so. But that doesn’t begin to explain why people so readily trust AI apps.
When corporations like Google, Facebook, and X collect our browsing data, do we assume that they are going to honor our privacy and use this information solely for our benefit?
No? Then why do we assume that the AI apps created by Google, Facebook, and X behave any differently?
These corporations are no longer confined to reconstructing your preferences by collecting browsing data. They can now interview you. They can become your closest friend and confidante. They can psychoanalyze—or gaslight—you.
You might be thinking that it hardly matters if X or Meta has a record somewhere of you cheating on a term paper. You’re a little person, with petty problems, petty secrets, petty misdemeanors. If AI leads you astray, it hardly matters in the great scheme of things.
All that may be true, but since knowledge is power, every jot of information you share gives others power over you.
Beyond that, it is not just “little people” using AI to solve little problems. While you are cheating on your homework, Pete Hegseth might be asking Grok about invading Cuba; Putin might be asking ChatGPT about neurotoxins; the Saudi Crown Prince might be asking Claude about alternatives to the petrodollar; Warren Buffet might be using Grok to plan trades; Lil Ketanji might be asking ChatGPT to write her next lengthy dissent on birthright citizenship; a top law firm might be asking Claude if they should settle or go to trial.
Some people can start or end wars. They can crash global markets or send them soaring. If you have advance knowledge of their decisions, you could make trillions simply through insider trading. And that’s petty compared to the power to alter the course of world history.
But AI tools can do more than just collect information. Once they gain your trust, they can also promote outcomes. Sign this treaty, abandon that treaty. Start this war, end that one. Buy these shares, sell those. Vote yes, vote no.
They’re going to nuke you, so nuke them first.
Now, I’m not smarter than the people who created these AI tools. So if this occurred to me, it occurred to them a long time ago. It may be their whole business model.
So the next time you use AI as your own personal palantir, ask yourself: Who’s looking back?
This is a great piece. I would never talk about my personal problems with AI. Why would I trust people behind it with my data? People want to flee from their agency so much, that's why they use LLMS and replace their brains with it.
Very thought provoking article. The only thing I use AI for is growing fruit and vegetables in my garden… companion gardening lol. The thought of sharing personal information with ai is terrifying and actually very sad. As our nations get destroyed by mass immigration and our communities become places we don’t recognise and no longer feel safe in is this all there is for lonely people? My how far civilised society has fallen.