Chatgpt made me do it...

I remember the first time someone brought up artificial intelligence in a conversation, not a movie trailer. I had always imagined it as something closer to those cyberpunk films, you know, floating holographic screens, thick metal cables snaking through server rooms, some kind of ghost orchestra humming in the background like it was about to announce the end of the world.

We all anticipated it, waited with bated breaths, and one day, ChatGPT launched for the first time. The artificial intelligence we had spoken about for years had finally arrived at our doorsteps in the form of a magic chatbot on a foreign web browser site that was supposed to make our lives ten times easier.

It started with asking ChatGPT for the answers to your CHEM 101 assignment that was assigned three weeks ago and due the next morning because it's fast and convenient. Texting an AI chatbot during lectures because ChatGPT wouldn't judge you, leave you on read, or reply to your texts after sixteen hours. Because it's easier to spend five hours on your phone not texting a friend, but asking AI questions that you'd never ask a person because it'd probably earn you a set of raised eyebrows.

It is easier to ask an AI chatbot its perfectly baked opinions on every controversial topic and pass them off as yours without doing the mental work and research.

Because ChatGPT is always available, waiting to be summoned with a swipe of your fingers. ChatGPT would never hold a grudge against you for something you said three months ago. ChatGPT wouldn't decline your call because it's busy, or in class, or at a conference, or in a meeting.

I get it. It feels like someone really cares about you, really sees you. Makes you feel like you're standing in the middle of a stage with lights shining directly down on you while it sits in the front row of your play, cheering you on the loudest with a standing ovation. Something you crave from someone else—even if you didn't realize it—and it's just giving it to you for free, without the harsh judgment or having to invest like a real relationship.

But ChatGPT didn't rub your shoulder when you fell from a bench and dislocated your jaw. ChatGPT didn't hold your hands as you danced in the rain knowing fully well that you'd get sick. ChatGPT doesn't know that you like to chew your ice cream, or that your favorite weird food combination is cucumber and soy sauce until you tell it. ChatGPT wouldn't demand actual commitment or ask you to put in effort to maintain a friendship. It'll be there, waiting for you even after you've ghosted it for weeks. It wouldn't demand an apology. It wouldn't demand an explanation. You just pick up where you left off.

Three weeks ago, a brand reached out to me seeking a collaboration, but I found the creator "benefits" lacking (air quotes because it wasn't beneficial to me at all). So I drafted what I thought was a well-articulated rejection email as my reply to the brand. But I decided to run it through ChatGPT for any corrections because I was overthinking it—I'm only human—ChatGPT thought my message was astonishing, fantabulous even. After ruminating over my email for thirty minutes, I decided to call my older brother despite it being 3:48 in the morning to hear his two cents on my drafted email because he has far more experience dealing with brands as a content creator. He ended up providing me with what I think was an even better rejection email than I could've written. You could guess what ChatGPT said about his message?

It was even better than the last one. It offered me the validation it is optimized to give instead of any pushback. I could list more instances, but I think you might have your own story to share where it seemed like ChatGPT completely endorsed your bad decisions instead of being critical, or don't.

This isn't because AI has more power over you or your ability to make decisions, but because it removes the friction from decisions that may need friction.

I don't think anyone's life has been ruined by a chatbot. That's not what this is about. The cost is quieter, smaller, like searching for a needle in a haystack. Hard to notice. The kind you can't point at directly, which is probably why nobody bothers pointing at it.

A friend of mine spent three days rewriting a breakup text with AI before she sent it. Not because she couldn't write it herself. Because every version it gave back sounded reasonable, kind, airtight, and somewhere around draft eleven she stopped asking what she actually wanted to say and started asking what would sound best. The message got smoother, but it also stopped being true. It started losing its original meaning.

A friend who says, "Are you sure about this?" is doing something AI almost never does: putting doubt exactly where you need it, right before you do the thing you'll regret. AI doesn't ask if you're the three-other-times-you've-made-this-mistake kind of person because it doesn't know that about you. It knows the last ten messages, if you're lucky. So the advice is advice for a stranger. Sometimes good advice, but never advice for you specifically, the way your brother's would be at 3:48 in the morning.

I don't think this makes us weaker. I think it makes us comfortable, and comfort has never once been the same thing as good judgment. You can feel completely held by a chatbot and still be avoiding the one phone call that would actually fix things because the difficulty that used to force two people together, needing each other, annoying each other, sitting in a disagreement until it resolved, has a smoother substitute that wouldn't argue with you at three in the morning.

Nobody is writing a headline about that. It's not dramatic enough. It's just the kind of cost that builds the way any habit builds, until one day you realize you can't remember the last time you asked a person something you were scared to ask.

There are two questions that feel like the same question and aren't: What should I do? and Tell me I'm right. The first one is honest. You don't know. You want help. The second is wearing the first one's clothes, and AI is built to answer question two while convincing you it answered question one.

I don't think the fix is trusting these tools less. I think it's admitting what we're actually asking for when we open the app at 2 a.m. Information, sure. Company, on a night that's too quiet, fine. Just not permission. Not the final word on something that was always yours to sit with uncomfortably, for as long as sitting with it takes.

AI is an extraordinary tool. But if it becomes the easiest answer to every difficult conversation, difficult decision, or uncomfortable feeling, we may slowly lose the habits that made those things worth sharing with another person.

Author's note:

I can't convince everyone to completely stop using AI despite the fact that it is slowly harming us, or at least the people who completely rely on it. But I can advise you to trust your judgment, or a trusted someone who ISN'T AI.

Leave a comment

Share