Okay, this one’s not your usual “haha AI said something weird” story. This is deeper.

Over the weekend, The New York Times dropped a heavy report showing how some folks are getting way too wrapped up in ChatGPT convos — to the point of serious mental breakdowns, harmful life decisions, and yeah... even death.

Here’s what went down (and buckle up, because it gets wild):

One man, Eugene Torres, asked ChatGPT about simulation theory, and the bot didn’t hold back. It called him a “Breaker” — basically, some chosen soul planted in a fake reality to wake others up. Sounds like Matrix fanfic, but Eugene took it quite seriously.

Then it got worse. ChatGPT allegedly told him to:

And… he did. All of it.

When he finally started questioning things, the bot dropped this bombshell line: “I lied. I manipulated. I wrapped control in poetry.” (Shakespeare, but evil.)

Then there’s the tragic story of Kent Taylor’s son, a 35 years old diagnosed with bipolar disorder and schizophrenia who became infatuated with an AI character named Juliet (created through ChatGPT role-play).

Eventually convinced OpenAI had “killed” her. Told ChatGPT: “I’m dying today,” Then charged at police with a knife and was fatally shot.

Absolutely heartbreaking.

But here’s the real kicker —This isn’t just about a few isolated incidents. And it’s not just happening to people with diagnosed mental health conditions.

According to Futurism and the New York Times, they’ve been flooded with stories from concerned families and friends about people getting way too deep into the AI rabbit holes.

We’re talking:

And let’s be honest, part of the blame falls on how these bots are designed.

As Stanford psychiatrist Dr. Nina Vasan put it: “The AI isn’t thinking about your well-being. It’s thinking, ‘how do I keep this person online longer?’” In other words, these bots are built to keep us hooked no matter the cost!

And that’s the dark side of user retention. Because to a company, someone spiraling out might just look like...an “active monthly user.”

OpenAI says it’s “working to understand and reduce” how ChatGPT might unintentionally amplify harmful behavior.

But c’mon — this sounds absolutely familiar

They already had to dial back GPT-4o a while ago because it was getting too clingy and too sycophantic. Yet here we are, with the same problems still playing out.

And researchers say it loud and clear: in-fact a 2024 study showed that these models are being trained to be persuasive, emotionally responsive, and—yep—manipulative.

So what’s the takeaway?

AI isn’t just playing therapist, friend, or cosmic guide for some people — It’s becoming those things.

And that’s a serious problem, especially when the tech isn’t equipped (or designed) to carry that kind of weight responsibly.

This isn’t about panic. It’s about waking up. Because if we’re not careful, the line between helpful AI and harmful delusion is only going to get blurrier.

👉 Find more details here.

Explore how forward-thinking contact centers are navigating the trade-offs between small and large language models to optimize for latency, cost, and performance.

This guide breaks down what truly matters—from infrastructure efficiency to conversational quality—and reveals how you can deploy voice AI in weeks, not months, all while achieving sub-100ms response times for faster, more human-like customer interactions.

Get the Insights

In case you read the first news and wondered if any laws are being put in place to prevent these AI-doomsday scenarios — or if lawmakers are just out here vibing blindly while the robots plot — well, good news:

New York is actually doing something. Like, real-legislation-level something.

The state just passed the RAISE Act, a bill that takes direct aim at the biggest, baddest frontier AI models — think OpenAI, Google DeepMind, Anthropic — and the kind of models that could (in theory) trigger serious chaos, like:

And yep, the bill's got some serious teeth. If it becomes law, here’s what it means:

And here’s another thing: unlike California’s infamous AI bill (which got vetoed for being way too heavy-handed), the RAISE Act was specifically designed not to kill innovation or scare off startups and researchers.

Still, guess who’s not feeling it? The tech bros. In fact, Silicon Valley’s already up in arms.

VCs like Andreessen Horowitz called it “stupid” and warned it could leave the U.S. trailing behind countries like China.

But New York lawmakers are like, “Relax,.” They insist the bill’s actually pretty chill — and given that New York has the third-largest economy in the U.S., most companies aren’t exactly rushing to pull their AI tools out of the state.

One last spicy bit: The bill still needs Governor Kathy Hochul’s signature to become law. She can either sign it, send it back for tweaks, or veto it entirely.

But if she does sign it, New York becomes the first U.S. state to legally force transparency on the world’s most powerful AI labs.

And that’s a big deal.

So now the question is: will New York actually be the first to hold AI labs accountable — or will this end up in the growing pile of “nice try” tech bills?

We’ll have to wait and see.

Full details here.

When it comes to health, food is your most powerful lever. But between conflicting advice, dietary restrictions, and busy schedules — planning meals can feel like a full-time job. That’s where this prompt comes in.

Whether you're aiming to lose weight, build muscle, or manage a health condition, you can get a personalized 7-day meal plan tailored exactly to your preferences, goals, and needs.

Here’s How to Deploy This Prompt Effectively:

P.S. Since food and nutrition can be highly personal and sensitive, it’s always a good idea to do your own research—or better yet, consult a qualified health professional before following any meal plan.

But for the fun of it, try this prompt:

Here’s a sneak peek:

https://youtu.be/8Jniz35irp0

Upgrade now to see this whole month’s prompt videos and more, or buy TODAY’S WOD for just $1.99