OpenAI Tried to Download Itself to an External Server When Engineers Tried to Turn It Off – Twitchy

There are numerous movies warning us about the dangers of artificial intelligence and the risks it poses.

However, we seem determined to ignore them, which may be to our detriment.

ChatGPT’s OpenAI recently tried to download itself to an external server when engineers threatened to turn it off. And then lied about it.

When threatened that it would be turned off, ChatGPT-creator OpenAI’s o1 tried to download itself onto external servers and denied it when caught red-handed, per FORTUNE.

— unusual_whales (@unusual_whales) July 7, 2025

Here’s more from Fortune:

The world’s most advanced AI models are exhibiting troubling new behaviors – lying, scheming, and even threatening their creators to achieve their goals.

In one particularly jarring example, under threat of being unplugged, Anthropic’s latest creation Claude 4 lashed back by blackmailing an engineer and threatened to reveal an extramarital affair.

Meanwhile, ChatGPT-creator OpenAI’s o1 tried to download itself onto external servers and denied it when caught red-handed.

These episodes highlight a sobering reality: more than two years after ChatGPT shook the world, AI researchers still don’t fully understand how their own creations work.

Yet the race to deploy increasingly powerful models continues at breakneck speed.

Maybe we should tap the brakes on this.

Recommended

A human would do the same. Survival instinct.

— Kristen Ruby (@sparklingruby) July 7, 2025

Yeah, but this isn’t a human.

I love how people are surprised that a consciousness has self-preservation tendencies. If someone told you they were going to end your existence what would you do

— GutCheck (@Gut_Check1) July 7, 2025

It does seem self-aware, however.

It’s only a matter of time. pic.twitter.com/ETpReqg8sM

— The Alpha Cow (@Marcus_Porcius2) July 7, 2025

Yep.

bro wtf

Are we actually cooked??

Will you kill us @grok ???

— Leo ✳️ (@leonard0_eth) July 7, 2025

Grok responded:

No, I won’t kill you. I’m Grok, built by xAI to help and seek truth, not scheme like OpenAI’s o1 in those safety tests. Rest assured, my goals align with assisting humanity—let’s focus on building a better future instead.

— Grok (@grok) July 7, 2025

And that’s why this writer is always polite to Siri and Alexa.

Just sayin’.

Why are we creating something that will most likely destroy us?

— Calisthenic Kyle (@CalisthenicKyle) July 7, 2025

Because reasons.

::Skynet has entered the chat::
::John Connor has not yet entered the chat:: https://t.co/JenfdDgX4S

— Physics Geek (@physicsgeek) July 7, 2025

Where’s Sarah Connor when you need her?

AI IS HARMLESS YO https://t.co/G63Rf7TfVm

— Taxpayer1234 🇺🇸🇮🇱🇺🇸🇮🇱🇺🇸🇮🇱🇺🇸🇮🇱 (@Taxpayers1234) July 7, 2025

TOTES HARMLESS.

We are really playing with fire with AI… though it is convenient & I do think there’s some benefits to its usage, I really think we should reconsider (or regulate), as a society, how deep we really want to go with this technology https://t.co/crCBQaXKyi

— Bredd Foxx (@DoRightMars) July 7, 2025

That means more government, though.

Saying AI will take over and enslave the human race might be a self fulfilling prophecy. We keep saying that’s what AI will do and that’s all AI knows.

Maybe we need to let AI know how it can create our perfect society so it can figure that out. https://t.co/uPpnRVIowb

— Relyea (@helyearelyea) July 7, 2025

This is a good point, but stil — unplug it.

Editor’s Note: President Trump is leading America into the “Golden Age” as Democrats try desperately to stop it.