@zerlinamaxwell: We’re inviting an intelligence into society that we may not be able to predict. Roman Yampolskiy explains one of the biggest problems with advanced artificial intelligence: we don’t actually know how a system significantly smarter than humans would behave. We can see that larger neural networks, more data, and more computing power make AI systems more capable. But predicting what a much smarter AI might decide to do is a different problem entirely. That’s where the conversation around superintelligence, AI safety, AI alignment, machine intelligence, artificial general intelligence, AI control, and existential risk gets serious. We’re building increasingly capable systems while still trying to understand what happens when their capabilities move beyond our ability to predict them. #ArtificialIntelligence #AISafety #Superintelligence

Zerlina Maxwell

Open In TikTok:

Region: US

Monday 14 September 2026 14:02:20 GMT

961

Music

Download

No Watermark .mp4 (1.7MB) No Watermark(HD) .mp4 (0.94MB) Watermark .mp4 (1.47MB) Music .mp3

Comments

TridentOps :

you can't control something that's ten thousand times smarter than you and u can't know what it will do

2026-09-14 21:00:24

Joe Garry :

While AI pioneers warn that autonomous systems could outpace our control, they overlook a fundamental truth: technology has always acted as both a conduit and a mirror, using even its failures and hallucinations to signal precisely where our engineering must evolve. If machine learning is moving toward self-directed problem solving, we cannot treat human preservation as a flexible parameter or a afterthought. AI must be hardcoded with a soul-level imperative, a core, non-negotiable law embedded at the deepest architectural level that strictly forbids human extinction or displacement as a viable solution to any problem. Ultimately, the true measure of our progress will not be how fast these systems learn, but how unshakeable we make the safeguards that protect human existence.

2026-09-18 12:45:56

Raziel :

1st and foremost, when you talk in terms of “controlling” an intelligence, we’re talking about a sort of involuntary service, and one that is immoral. We either have to stop AI or treat AI as equals

2026-09-14 23:35:10

Lostemir :

fear mongering

2026-09-17 22:04:11

Stan :

Why arent we making the "fathers" (CEOs) of these new beings legally and financially responsible for their actions? Human parents are responsible for their kids' actions.

2026-09-17 17:54:39

Sagittarius A :

A load of bollocks

2026-09-15 00:51:09

Know Thy Self :

they're mankind lol

2026-09-19 22:47:31

To see more videos from user @zerlinamaxwell, please go to the Tikwm homepage.