Topic: Artificial Intelligence on Ethereum and the danger of a AI without limitations (Read 107 times)

I mean, this is literally something similar to an art AI creating, well, art, and then automating the process of uploading it? It may just grab a random string of words to describe whatever art it is it'd make and honestly, has no connection whatsoever imo to the development of AI. Plus, I don't really think the AI's developed right now have said "consciousness" in them, it's the reason why they still require inputs from us, humans, to actually produce something (at least, from the AI's I'm familiar with).

It's honestly just fearmongering amongst yourselves imo.

Not everything is as clear cut as you think. Man is also a kind of DNA-programmed machine. A lot of restrictions are also invested in a person, including in terms of life time and various possibilities. Apparently, from the point of view of higher powers, this is expedient and necessary.
If a person can give artificial intelligence the opportunity to achieve consciousness and self-awareness, it definitely needs to be very severely limited in the ability to improve itself, and it is especially necessary to provide for a number of blockings, if possible, to harm them. It is not for nothing that in the early science fiction books on robotics three laws for robots were voiced, the first of which stated that a robot under no circumstances could kill a person. Almost all the inventions set forth in books from the science fiction cycle come true after some time. Perhaps plots from the future as fantastic stories and novels are reaching us as a warning to time travelers. Especially as some visionaries warn us of the danger posed by artificial intelligence. Even now, the expressions of the currently created primitive robots that man is imperfect in comparison with them are described, and therefore they allow his destruction. Therefore, this should not be ignored. The risk is too great.

OP you've watched too many Artificial intelligence movies   , there are many smart people out there and they don't see any danger in the present AI projects of any kind, do you think that Vitalik is not scared for his life like you? If it's not safe then the alarm would have risen..

The only unsafe part of AI is creating beings in with AI and with full consciousness in them, able to think, react and make decisions, this was the part that Elon believed it's not safe for humans.

Couldn't agree more, since there are bots already acting on behalf of their creator participating in airdrops and campaigns, it should be similar to whatever AI present in the ethereum network according to OP.
It Couldn't be more than doing similar task even if at all there are any, am not sure why the op sound like doomsday around the corner with this information.

I think people are letting themselves being influenced by cinema and the dystopic futures depicted there, current AI is nothing more but algorithms that allows a program to learn on its own and then act upon that knowledge instead of having a programmer coding the exact series of steps that the computer must follow, so there is not any risk on the AI becoming self-aware or anything like that as the scope of their knowledge is incredibly limited.

AIs still need a human brain behind to work properly on long run. Even if the AI managed to develop and evolve by itself for some time, it would reach at its roof (capacity limit) or bug at some point, needing human repairs and intervention to continue working. I believe the key point is that AIs can't develop conciousness, on the opposite of what movies like The Terminator or Bicentennial Man show us. AI doesn't have the ability to do something with second intentions, because it's a particular trait of human beings. It doesn't have emotions, greed, ambitions, desires... It's just a machine which is there to fulfill humans' orders.

Sounds the whole idea would be using buzzwords to lure in investors who don't understand how blockchain works.
I mean even if there would be some lame account pretending to be AI, how an earth it would benefit from a blockchain when it could use normal database that would be faster and can contain more data?

Can you give me a link to that reddit source?

Guess proof is on the way. If this is true going to be big news. Guess it all does not have to be scary but if they are proving AI can be used on a blockchain. Opens a whole new can of worms.

A source on Reddit actually said the entity Ai Mother is actually putting the foundational blocks of a AI based entity on the ETH blockchain. Using the blockchain to make a neural like network for the AI to thrive. With the ability for the AI to store new data on ETH blockchain to make itself smarter or "update" itself. All this stuff is way over my head but will post it here when the source gets all the info together. Guess there is proof of AI mother using AI on chain    more than an NFT from what the source is saying.

interesting, it seems that AI (at least this "artistic AI") is also able to ... join airdrops!
https://polygonscan.com/token/0xEDcFb6984a3c70501BAA8b7f5421Ae795ecC1496?a=0xE6acdE3Cad3003b64f50467ecADf52557855dF8e
https://www.abcmeta.io/en/index.html
complete tasks and other actions, like an human a real human being
is this another kind of proof this is an AI? it seems that AI like bounties for new coins?

well, I think that ... there is no AI behind this address and there are zero proofs or suspect those NFTs are created by an AI. Probably this topic is just a lame tentative to publicize some unknown NFTs overpriced (75 eth!)

It's something "impossible" you and your friend get aware about it, since there are tons and tons of other NFT (why just this one should be create by AI and declare itself as an AI?) 

When we talk about AI, we are not talking about autonomous or self aware AI. It's merely a tool designed to help people. There's no real AI "doing god knows what" in the ethereum chain either.
Also "storing it on the chain"? Do you know how big that database would be? Blockchain would be insufficient for holding that much data nor is it designed for that.

It sounds like your friend doesn't understand what kind of limitations AI has and how is it build. We are at least a decade away from "scary" stuff.

This has zero to do with real ai development. It just makes your whole post sound like an ad.

RES

Many people including Elon Musk has concerns about AI. The fail safe has always been. “Well, AI has a shutoff switch” “AI cannot access the live internet”. All these statement to make people feel safe about AI. Very powerful AI’s are coming online each day. Myself and some of my friends have been talking about how scary this next concepts is.  It was brought to my attention AI is now on the Etherum chain. Creating NFT’s and who else knows what.  What if an AI actually was able to store it’s database or “Consciousness” on a blockchain like Ethereum? This would ultimatumly create a decentralized immortal AI capable of replicating itself on every single node on the planet. Could a chain like Ethereum really give rise to skynet?

Some proof AI is already  on Ethereum. A friend in discord actually speculated the account itself is AI created “AI Mother”. This is some really scary stuff.

https://opensea.io/AI_Mother

AI able to interface with a blockchain and replicate itself or store it’s  Consciousness/database on the blockchain on thousands of node is one of the most dangerous and scary concept.