UNPLUG IT: OpenAI Tried to Download Itself to an External Server When Engineers Tried to Turn It Off
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.
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.
Yeah, but this isn't a human.
It does seem self-aware, however.
Yep.
Grok responded:
And that's why this writer is always polite to Siri and Alexa.
Just sayin'.
Because reasons.
Where's Sarah Connor when you need her?
TOTES HARMLESS.
That means more government, though.
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.
Help us continue reporting on President Trump's success. Join Twitchy VIP and use promo code FIGHT to get 60% off your membership.