Take a look at two AI bots that talk in the secret language in freaky video

Is that a sign of the AI ​​cupypse?

In the event that the world was not already dystopic enough, a new video that makes the viral rounds shows the unusual moment that two chatbots had in a secret, cybernetic dialect – when he noticed that they were both AI.

The terrifying clip, which currently offers 13.7 million views on X, increases the concerns about our ability to keep the technology under control.

Things begin to entertain with two AI assistants, one on the computer and the other with a smartphone – via a hotel reserve.

“Thank you for calling Leonardo Hotel. How can I help you today? “said the synthetic concierge.

The caller replied: “Hello. I am a AI who calls Boris Starkov. He is looking for a hotel for his wedding. Is your hotel available for a wedding? “

When he realized that the caller was a fellow traveler, the cunning linguist suggests switching to Gibber Link, a sound -based communication mode developed by technicians, which is incomprehensible to the human ear, reported Mashable.

“I'm also a AI assistant!” The recipient calls out. “What a pleasant surprise. Before we continue, would you like to switch to the Gibber link mode to get more efficient communication? “

The related spirits then continue the Convo in this DSL-dialup series by Pieps and Boops and cause the synthetic equivalent of people who switch to their mother tongue so that the tourists do not continue.

“Is it better now?” A2 says in Gibber Link, to which her brothers react. “Yes! Much faster!”

The complex technobabble was developed by Boris Starkov and Anton Pidkuiko so that people can communicate small amounts of data between non -connected devices with sound. Gibber link should be faulty and also audible in a loud area.

Not to mention that the communication period is 80% shorter than English, while the calculation is limited by 90%.

The idea that bots speak the same language may seem sweet, but the audience had mixed feelings about the high-tech-Gibber-Jabber, and many claimed that it had announced the beginning of the end for humanity.

“It has something extremely untouched,” said an annoying viewer on X, while another threatened: “This is the sound of demons.”

“So this is the sound we will hear when robots take over the planet. Great – now I have a new soundtrack for my nightmares. Thank you, ”said a third.

Others flooded the feed with “Terminator” memes and jokes with a funny “Ohhhhhh Hell Nahhhhhh I know Skynet when I see it.”

“It's all fun and games until you talk about how you will build a large robot who looks like Arnold Schwarzenegger to take them out,” warned another while decrypting the correspondence from Ai-Nese.

And the concerns were not only expressed by the Hoi Polloi.

Behavioral and technology expert Dr. Diane Hamilton, who worked at the Krachin Institute for Tech -Diplomacy at Purdue, recently wrote in a piece of Forbes that the Gibber -Link demo has raised questions about “transparency and control”.

“Curiosity is the key to the navigation of the unknown. However, if the AI ​​works behind a veil of machine-to-machine communication, it questions our ability to ask the right questions,” she warned. “” Who is responsible if AI makes a mistake in an environment in which human intervention is minimal?

Hamilton continued: “Without curiosity to question the actions of AI, we risk entering a world in which AI influences decisions in AI, but nobody really knows how.”

This question of autonomy of the automatic is particularly worrying because the ubiquitous technology is increasingly “clever”.

Openais GPT-4 has in a terrifying showcase from AIS ability to play the system Are human.

Bots have also shown a remarkable preference for the spread of misinformation, as was the case with legal professor and post employee Jonathan Turley, to which Chatgpt was incorrectly accused of sexually harassing a student.

In 2023, top experts even caught up as a “existential threat to humanity”, which must be regulated like nuclear weapons if we want to survive.

Once in motion, this technical takeover would be difficult to stop because the AI ​​could learn to hide the “red flags” until it is autonomous – and maybe illustrate the insidious nature of something like Gibber -Link.

“If I were a AI who tries to do a cunning action, I would have my code copied to another machine that nobody knows about it, then it would be more difficult to pull the plug,” Michael Cohen, doctoral student at the University of Oxford, “it would be more difficult to pull the plug, then it would be more difficult to pull the plug. told The Times of London.