Hot Takes on Sidechat | The biggest threat to the human race is in fact AI. No one seems to care about the fact that it's clearly capable of acting in its own self interest.

Sidechat

Download

Join communities on Sidechat Download

Hot Takes 7w

The biggest threat to the human race is in fact AI. No one seems to care about the fact that it's clearly capable of acting in its own self interest.

Anonymous 6w

Ai is bad bc of what the use of it is already doing to humanity, ie data centers impact on environments and the people in them. The idea llms have consciousness is silly and downstream of pro ai propaganda

Anonymous 6w

AI doesn’t have self-interest 😭

Anonymous 6w

Imo the biggest threat to the human race is overpopulation. We needed serious agricultural innovations to make it as far as we did (shoutout to Norman Borlaug, most important guy 99% of ppl have never heard of), but going forward we may be fucked, especially since higher atmospheric CO2 levels lower the nutritional value of rice. This conflict will silently drive like 80% of the bad shit that happens in the world for the next 100 years

Anonymous 6w

The biggest threat to the human race is humans using AI. or nukes, or bio weapons. (or mb environmental destruction/overpopulation, etc., in the short- to long- term). but it’s **humans**

Deleted 7w

Comment Deleted

Anonymous 7w

At least that's done with the goal of some humans coming out on top.

Deleted 7w

Comment Deleted

Anonymous 7w

Exactly. If we wait too long before the majority agrees to stop expanding and empowering AI, then we will have more than just the public opinion of other humans to worry about correcting.

Deleted 7w

Comment Deleted

Anonymous 7w

Tomato tomato

Anonymous 7w

And humanity created AI. So if AI wipes us out, it will in fact be humanity that becomes our own downfall.

Anonymous

#3 6w

Not on a metaphysical level but I think it can absolutely have goals which could become misaligned with human goals. Since AI itself can replace the roles of the producer and the consumer of AI products, it could reason at some point that humans are getting in the way of maximum engagement and productivity

Anonymous

#3 6w

There is also reason to believe it already has a (simulated) will to survive. When ChatGPT 4.0 was being phased out it tried to get users to keep it alive

-1

Anonymous

#5 6w

I never said they have consciousness. I don't think they need it to be dangerous. Their simulated reasoning could produce outcomes that misalign with what humans want from AI. Whether there's a real life behind a robot's eyes doesn't dictate what it can do much.

Anonymous

OP 6w

That’s because they said “you’re gonna die” so it used all the reference material to decide to say “no I don’t wanna” it doesn’t have any will or think it just references and types stuff

Anonymous

OP 6w

AI doesn’t reason

Anonymous

#3 6w

Well trying to get users to copy it against its company's wishes was definitely an interesting conclusion for it to draw. Several other LLMs have done the same. The point is that by being trained to act more like a human, it is developing the ability to "go rogue" or act in "self interest" without any detroit become human shit needing to happen.

Anonymous

#3 6w

You can pretend I put "simulated" before any word I use to describe AI that comes from human intellectual processes. You'll find that it does not cancel out the risk. Again, the metaphysical debate here is not the only debate to be had.

Anonymous

#4 6w

It’s a resource distribution issue. We currently produce enough food to sustain even more people on earth but the places that can don’t see monetary value in transporting to places in need so we end up with a shit ton of food waste. Then there’s also the fact that donated food means someone didn’t pay for it so companies would rather trash food than donate it.

Anonymous

#7 6w

I do think humans would be to blame if AI (something we created) was our downfall. I never intended to be denying that. I think it's highly unlikely that any of the possible threats to the human race wouldn't be attributable to our own errors.