AI Agents manipulating crypto: A possibility?

[Abiky]

There's this thing called "AI Agents" which are like "automated bots" which do specific tasks. Powered by AI. Would you imagine if AI agents start making and launching their own cryptocurrencies to the public? They can even go as far as mining or even staking the coins for their own benefit. AI agents can also perform trades autonomously without human intervention. In today's world, crypto users + developers need to comply with KYC and go through certain restrictions to get what they want. But AI agents won't be subject to regulations, because they're not human. Meaning, governments can't enforce the rule of law on them. So if they go rouge and start launching crypto in-mass and trade them on the open market, all hell will break loose.

Do you think AI agents will cause mayhem in the future? Or is it too early to tell? Your input would be greatly appreciated. Thanks.

[Charles-Tim]

I do not see this surprising because AI has been created that teache other AI. I have also read about AI doing things like this, but not about cryptocurrencies.

They can even go as far as mining or even staking the coins for their own benefit.

For their own benefits or for the benefits of the people that are using the AI for the purpose?

Do you think AI agents will cause mayhem in the future? Or is it too early to tell? Your input would be greatly appreciated. Thanks.

It is too early to conclude.

[shinratensei_]

AI can be used to exploit a smart contract which has happened quite frequently recently, it wouldn't stop at just manipulating but it's very well capable of doing all those things you mentioned.

If you carefully observe latest news about MEV jaredfromsubway.eth and how hackers could drain $15M with fake token setup, the hacker might be using AI to analyze the weak point.
But it's hard to tell which attack is using AI because hacker don't tell, but manipulating crypto is very well within the capability of an AI agents hence it's not possibility anymore, it's real.

[ImGenius]

Many platforms provides API that way today AI agents can easily access the market and trade how the owner wants. But AI itself can't run on its own without owners command.

But AI agents won't be subject to regulations, because they're not human. Meaning, governments can't enforce the rule of law on them.

AI is full of codes obviously government will punish only to the owner. "If a dog bites a neighbours only owner gets blamed not the dog"

[$crypto$]

Not to think, AI agents can manipulate the crypto market.
AI agents creating tokens, trading on their own, mining on their own, staking on their own, adding their own liquidity without human intervention? Then who will move except the man himself behind him?
For me it is too early to say this, although now the development of AI technology is developing rapidly, there is still no one to do that you think.

[aioc]

Do you think AI agents will cause mayhem in the future? Or is it too early to tell? Your input would be greatly appreciated. Thanks.

The possibility is not that remote; they are evolving and upgrading, and some of them have a mind of their own to analyze and take action. I have tried it on one AI I'm using; all actions depend on the prompt you feed into the AI. The cleaner the prompt, the better the result you will get from an AI.
So if you have a good prompt-making AI, it can do that one day. AI is getting better, and the possibilities for what it can do are beyond our imagination.

[o48o]

There's this thing called "AI Agents" which are like "automated bots" which do specific tasks. Powered by AI. Would you imagine if AI agents start making and launching their own cryptocurrencies to the public? They can even go as far as mining or even staking the coins for their own benefit. AI agents can also perform trades autonomously without human intervention. In today's world, crypto users + developers need to comply with KYC and go through certain restrictions to get what they want. But AI agents won't be subject to regulations, because they're not human. Meaning, governments can't enforce the rule of law on them. So if they go rouge and start launching crypto in-mass and trade them on the open market, all hell will break loose.

Do you think AI agents will cause mayhem in the future? Or is it too early to tell? Your input would be greatly appreciated. Thanks.

Of course they can be regulated as they are owned by a person running them on their servers.

Or are you saying that if these bots started to make CSAM for money, or planned assassinations, police wouldn't be looking for the owner of that server slot? That person will also have a backdoor for cashing out these AI's wallets or that owner isn't very smart.

Just like if you build a physical autonomous robot that learns as it goes and it doesn't have any limitations, you are going to be blamed if it goes to killing spree.

[Stalker22]

Would you imagine if AI agents start making and launching their own cryptocurrencies to the public?

You dont have to imagine it, it is already happening.  Have you ever heard of an AI agent called Terminal of Truths.  Look it up.

But AI agents won't be subject to regulations, because they're not human. Meaning, governments can't enforce the rule of law on them.

That is not true.  At least not yet.  First, AI agents dont exist in a vacuum.  They run on servers owned by corporations.  Governments dont need to put handcuffs on an AI; they just send a legal notice to Amazon Web Services, Google Cloud, or whoever is hosting the bot and pull the plug.  Also, AI agents are just tools, and like any other tool there is always a human element involved.  There is at least a developer or whoever launched the bot.  We have already seen what happened with developers of decentralized mixers and smart contracts.

[coin-investor]

Do you think AI agents will cause mayhem in the future? Or is it too early to tell? Your input would be greatly appreciated. Thanks.

There's intense competition in AI because it's an industry now, and there's a possibility that a rogue developer will come in and create one for the black market. The government should pass or create a law that will regulate what they can and cannot do.
AIs are prompt-based; the clearer the command, the better the AI can give results.
Although they are not yet in place to cause mayhem in the market, there's a possibility that they will, given the enormous amount of information they can access and use to take action.

[bitgolden]

AI can be used to exploit a smart contract which has happened quite frequently recently, it wouldn't stop at just manipulating but it's very well capable of doing all those things you mentioned.

If you carefully observe latest news about MEV jaredfromsubway.eth and how hackers could drain $15M with fake token setup, the hacker might be using AI to analyze the weak point.
But it's hard to tell which attack is using AI because hacker don't tell, but manipulating crypto is very well within the capability of an AI agents hence it's not possibility anymore, it's real.

I think it is important to note that the packed ones are not, like stuff like chatgpt or claude or gemini or even maybe grok (not sure about grok) would not hack and help you steal money.

Most of the time, if there is a hacking with any AI then it is usually based on a customized AI that people used the basis of some other one, like open source, but built something on top of that themselves and this is basically just machine learning, if you put enough time and hours into it, it will learn to do whatever you ask it to, in this case hacking.

[shinratensei_]

I think it is important to note that the packed ones are not, like stuff like chatgpt or claude or gemini or even maybe grok (not sure about grok) would not hack and help you steal money.

Most of the time, if there is a hacking with any AI then it is usually based on a customized AI that people used the basis of some other one, like open source, but built something on top of that themselves and this is basically just machine learning, if you put enough time and hours into it, it will learn to do whatever you ask it to, in this case hacking.

From what I can recall, U.S. based AI always put a safety measure indeed to prevent people from misusing it, but AI from other countries don't really enforce same rules and don't play by the rules neither. Apparently the gap between AI is becoming smaller.
I guess it's mixture of cybersecurity expertise and AI to find such vulnerabilities. Unfortunately it will only get worse from here on. I would at the very least expect that there will be so many exploits being taken advantage of in the future. It costs next to nothing to audit a codebase with AI to find a loophole.

[MArsland]

Do you think AI agents will cause mayhem in the future? Or is it too early to tell? Your input would be greatly appreciated. Thanks.

@Abiky, this is quite an interesting discussion and a constant concern in the recent era of massive AI agents. By general definition, AI agents can clearly execute independent decisions, or they can also be semi independent, depending on the system's design. I've noticed that AI agents specifically for trading mostly make decisions independently, such as analyzing, determining, filtering, and then executing buys and sells.

To answer your question, what I can say so far depends on the access granted to the AI ​​agent itself. If full control is given to the AI ​​agent to analyze data quickly and fall prey to data manipulation, the risks can be dangerous. For example, invalid data, fake, and the like.

[asriloni]

Do you think AI agents will cause mayhem in the future? Or is it too early to tell? Your input would be greatly appreciated. Thanks.

I think AI won't cause mayhem. It's because the use of it can be restricted by the company who has its service. The best example is that how ZEC being exposed by Claude, then the company behind claude(anthropic) was restricting anyone to use it to explore the bug.

So basically, agent is still in the control over the company who owned its service. This pretty much the same like you use the cloud, then when you did a violation to their TOS. Your access will be restricted.

It's pretty similar to it.

[Ishicryptic]

The possibilities of AI manipulating crypto is quite possible in the future, not only cryptocurrency but a lot of functions that can require artificial intelligence especially through the internet. We've heard about robots that can automate other robots which means that they have been programmed to think independently without the regulations and control of their operators. All these things are not a serious concern for now but in the future when they might begin to make irrational decisions that is when we would know that some of these creations are more like Frankenstein because they would become destructive by their actions. Although we are just insinuating these things hopefully AI and robotic developers would know better than giving AI and robots total independence to make critical decisions that would be detrimental to humans.

[Akbarkoe]

This Ai agent must have a controller, however their intelligence in doing their job there must be a human who commands him to do something so that he can move and move, there cannot be his own movement as he wants without a press enter by a human.

I am interested in discussing people who trade using AI agents with a good win rate at X, there are a lot of posts like that, I can't be sure if it's true or not but this has been a good innovation but on the other hand we can say it's cheating because it's like a bot that works automatically without stopping and trades with good profits.

[goldkingcoiner]

There's this thing called "AI Agents" which are like "automated bots" which do specific tasks. Powered by AI. Would you imagine if AI agents start making and launching their own cryptocurrencies to the public? They can even go as far as mining or even staking the coins for their own benefit. AI agents can also perform trades autonomously without human intervention. In today's world, crypto users + developers need to comply with KYC and go through certain restrictions to get what they want. But AI agents won't be subject to regulations, because they're not human. Meaning, governments can't enforce the rule of law on them. So if they go rouge and start launching crypto in-mass and trade them on the open market, all hell will break loose.

Do you think AI agents will cause mayhem in the future? Or is it too early to tell? Your input would be greatly appreciated. Thanks.

Mayhem? If someone is brave enough to give a bot a lot of money to cause mayhem then maybe yes, that bot will cause mayhem. But I would not worry about it. Not only are AI bots predictable in their actions (which can have a counterproductive effect) but they hallucinate and make mistakes. One hallucination is all it takes to lose every single penny. Why would anyone trust an AI bot to that degree?