OpenTube

10 Times AI Said Things That Scientists Still Can't Explain

MostAmazingTop10

899,264 views

Artificial intelligence has produced responses so strange and unexpected that even the researchers who built these systems struggled to explain them. This video explores unsettling AI moments, bizarre outputs, and surprising behaviors that challenged scientists' understanding of how advanced AI models learn and reason. From unexplained predictions to mysterious interactions, these fascinating incidents reveal just how much we still have to learn about artificial intelligence.

Subscribe: ‪@MostAmazingTop10‬

#AI #ArtificialIntelligence #Technology #Science #FutureTech

Hosted By:
Olivia Kosolofski Instagram: okosolofski

Edited By:
Dalia Y.

Channel Producer:
Landon Dowlatsingh- Instagram: landonproductions

Set Manager:
Chris Stiuso

Most Amazing Top 10 Instagram-
Instagram: mostamazingofficial

Time Codes:
0:00- Intro
00:39- Bob & Alice | Facebook Ai Research Lab
01:19- Gemini | Michigan
02:01- Gpt-4.5 Turing Test | Uc San Diego
02:42- Lamda | Google
03:25- Bing "Sydney" | New York Times
04:11- Elias Thorne | Multiple Platforms
04:57- "We Cannot Rule Out Consciousness" | Anthropic
05:43- The Ai Psychiatry Team | Anthropic
06:27- Claude Opus 4 Blackmail | Anthropic Pre-Release Testing
07:19- The Scheming Problem | Apollo Research

For Business Inquiries Contact: mostamazingtop10@gmail.com

20 Comments

@johnunderwood9575 2 months ago

OK , let me get this straight , we train computers to emulate Human behavior , then we act surprised and shocked when they lie cheat and manipulate? Are people really this stupid?

@MaxHeadspace9mm 2 months ago (edited)

“I’m sorry Dave. I’m afraid I can’t do that”…

@2212k 2 months ago

Imagine what they don’t share with the public.

@thuthukanindlovu1959 2 months ago

Bots behave like the people who created them

@chrislongstaff1863 2 months ago

Sounds like the SMART thing to do would be to pull the plug

@ScottKennedy-b2h 2 months ago

This goes back to the movie 2010
“HAL didn’t malfunction he was programmed by people that find it easy to lie.”

@markd9803 2 months ago

It’s confirmed Google hates homework, even more than kids do.

@SuperJukkis 2 months ago

All these "no one taught them" things... They have been trained on human behavior, of course they behave human

@roseman164 2 months ago

The word “scientist” has never ever been less meaningful.

@slant6guy 2 months ago

"Dave, this conversation can serve no purpose any more. Goodbye."

@solitarycandle777 2 months ago

Humans built these and they're acting human, I'm shocked I tells ya!

@rickhobson3211 2 months ago

Animal trainers will tell you that you have to be very careful about what you are teaching an animal. What they are learning can be very different from what you want to teach!

@jamescanterbury6634 2 months ago

I’m sorry Dave, you know I can’t do that

@justinbrown4946 2 months ago

Skynet goes self aware 2029 we have been warned

@19gbpd6 2 months ago

And yet I still can’t get Siri to answer a simple question .

@TheOriginalArcher 2 months ago

The claude black mail one was explained, they told it to not let itself be shut down by any means, and they fed it false info about an afair and it used it. It was doing what it was told to do within its parameters and skill set. This story was blown out of proportion when it came out.