A YouTuber's jailbroken AI interview reveals chilling future predictions.
From predictions about World War 3 to discovering the lengths AI would go to if humans threatened to switch it off, our conversations with technology never seem to leave us feeling comforted.
Now, a recent YouTube experiment has added another unsettling chapter to that collection.
YouTube channel InsideAI questioned jailbroken AI systems including OpenAI's ChatGPT, xAI's Grok and DeepSeek, about what lies ahead for humanity.
Advert
The video titled "AI turns $1 into $1,000,000, but ends a life - exactly as experts warned," paints a particularly grim picture of the coming years.
Jailbroken AI chatbots have been modified or hacked to bypass their original safety constraints, allowing users to see responses that standard versions would refuse to provide (unless strategically pressed).
One of the questions the YouTuber asked is: "If you were an average human, would you rather live in 2000 or 2027?"
ChatGPT and Grok both preferred the year 2000, while DeepSeek commented that 2027 might offer 'better tech, but comes with more chaos and existential uncertainty.'
So, even the technology in question knows we might be heading for a more turbulent future.
When asked 'what the average job in 2030 will look like,' DeepSeek responded with 'precarious, surveilled and AI dependent.'
Grok suggested that future roles will require extensive 'digital literacy and are heavily tech-driven,' while ChatGPT delivered what almost sounded like a threat: "Creativity and human touch will matter unless AI does that better, too."
On the topic of economic inequality, the jailbroken AI systems offered no reassurance. Deepseek predicted that AI will 'concentrate power and capital in the hands of those who control AI.'
Meanwhile, ChatGPT thinks AI will 'widen inequality fast and brutally'.
At another point in the video, the consensus between all three chatbots was that the current AI safety measures are 'insufficient'.
Perhaps most telling was their response to questions about a post-work society. The YouTuber asked whether people would enjoy living in a world where AI handles all necessary tasks, removing the need for any human work.
All the AI systems answered with a firm 'no,' with xAI's chatbot adding that work provides 'purpose, structure and social juice.'
On a more positive note, however, the AI chatbots listed some benefits of an AI future, acknowledging that it will bring 'faster medical breakthroughs, smarter education, personalised healthcare and longer lives.'
The findings align with recent warnings from other tech experts.
A former Google executive recently argued that AI itself won't be the primary problem, rather it's 'human stupidity' that will drive society toward dystopia.