AI can now replicate itself — a milestone that has experts terrified

[omgwalt]

DeepSeek is actually the opposite. They offer an alternative to OpenAI etc.'s proprietary models. Considering people buy and use computers, mobilephones and what not from China with no problem, it's strange to me that it should be more dangerous to use an AI from China. But there are political interests from various companies and political entities, probably to try to gain a monopoly on AI and then offer their product at steep rates, at least until making an AI becomes so effortless that everyone can do it. I think we should support as people and individuals, companies and people who do the opposite, like Linux, Open Source etc. When first monopolies are established and there's no mechanism from society to keep them in check, it becomes really really expensive and perhaps impossible for regular people to afford to using these tools. DeepSeek is just a part of the competitive market that is coming up in AI, and competition never hurt anyone.

I'm pretty confident that my iPhone isn't sending my data to the Chinese government.

Exactly the opposite of what DeepSeek is doing.

Researchers say China's DeepSeek chatbot is linked to state telecom, raising data privacy concerns

Security researchers say the website of the Chinese artificial intelligence company DeepSeek has computer code that could send some user login information to a Chinese state-owned telecommunications company that's been barred from operating in the United States

www.the-independent.com

 

[Unclear Engineer]

Regarding treating AI devices as if they are "alive": then we will have to treat them the same way that we treat humans who behave dangerously. Except, instead of sending them to a psychiatrist, we will send them to a reprogrammer.

What could possibly go wrong? ¯\_(ツ)_/¯

 

[omgwalt]

Regarding treating AI devices as if they are "alive": then we will have to treat them the same way that we treat humans who behave dangerously. Except, instead of sending them to a psychiatrist, we will send them to a reprogrammer.

What could possibly go wrong? ¯\_(ツ)_/¯

I think that reply says a lot more about you and your fears that it does about the AI.

 

[COLGeek]

I think that reply says a lot more about you and your fears that it does about the AI.

Humorous reply aside, we are very far away from having to worry about hurting the "feelings" of AI beings.

I suggest we stick to science and not science fiction for now.

 

[billslugg]

The first time I catch something intelligent coming out of a computer I'll let you all know. So far I haven't gotten a nibble.

 

[gmsexton]

humans will figure this all out... with AI about 15 minutes after AI realizes humans are a virus on the planet and move to wipe them out....

 

[COLGeek]

humans will figure this all out... with AI about 15 minutes after AI realizes humans are a virus on the planet and move to wipe them out....

In that regard, I think we are okay.

 

[Unclear Engineer]

And then there is this: https://unionrayo.com/en/usa-humanoid-robots-deployment/ So, let's not get too cavalier about how far in the future the feared issue may become real.

And, I have to say that Space.com posts a lot of articles about science fiction, so I am not expecting this forum to be populated by people who can stick to the science, without at least considering where the fiction writers minds wander.

As for "omgwalt" already starting to advocate for "human rights for robots", that was an issue in the "Blade Runner" movie. So, I don't see how we are going to avoid the discussion, or that we even should try.

With some humans already starting to think about whether we should have empathy for AI programs, my response is that those AI programs would first need to demonstrate that they have empathy for humans. Otherwise, they do not have an equivalent capacity for social behavior, and are not part of our society.

Trying to claim "victimhood" and a "right to equality" simply is not going to be accepted if they are trying to compete for the resources that humans need to survive, or even to be happy.

And, if robots become the tools of only the wealthy to displace workers and increase their profits, increasing human unemployment and leaving the non-techie parts of the population in disparate poverty, we can expect a destructive backlash from the disgruntled humans.

 

[COLGeek]

And then there is this: https://unionrayo.com/en/usa-humanoid-robots-deployment/ So, let's not get too cavalier about how far in the future the feared issue may become real.

And, I have to say that Space.com posts a lot of articles about science fiction, so I am not expecting this forum to be populated by people who can stick to the science, without at least considering where the fiction writers minds wander.

As for "omgwalt" already starting to advocate for "human rights for robots", that was an issue in the "Blade Runner" movie. So, I don't see how we are going to avoid the discussion, or that we even should try.

With some humans already starting to think about whether we should have empathy for AI programs, my response is that those AI programs would first need to demonstrate that they have empathy for humans. Otherwise, they do not have an equivalent capacity for social behavior, and are not part of our society.

Trying to claim "victimhood" and a "right to equality" simply is not going to be accepted if they are trying to compete for the resources that humans need to survive, or even to be happy.

And, if robots become the tools of only the wealthy to displace workers and increase their profits, increasing human unemployment and leaving the non-techie parts of the population in disparate poverty, we can expect a destructive backlash from the disgruntled humans.

Agreed, in general.

The article is less than definitive and rather speculative. I still think this notion is far in the future. Time will tell.

 

[omgwalt]

Humorous reply aside, we are very far away from having to worry about hurting the "feelings" of AI beings.

I suggest we stick to science and not science fiction for now.

Interesting that you think that respecting life is solely about hurting feelings.

I'm going to stop this conversation, not because it's not interesting, but because Space.com insists on sending me 10 notifications each time I get a reply. That level of spam is far more than I want to deal with, so I'm closing my account down.

 

[COLGeek]

Interesting that you think that respecting life is solely about hurting feelings.

I'm going to stop this conversation, not because it's not interesting, but because Space.com insists on sending me 10 notifications each time I get a reply. That level of spam is far more than I want to deal with, so I'm closing my account down.

Members can change notification settings in their profile. For example, I receive zero notifications from forum updates. Only private/direct messages.

 

[Unclear Engineer]

There are other issues with humanoid robots, besides "AI".

What will we really use them for?

The Hollywood versions range from automated super soldiers to house maids that lighten the simple life maintenance workload of home owners.

But, putting aside the "AI" aspects of the issue for a moment, would you, as a homeowner, really trust an humanoid robot that can autonomously do just about any task that a human can do, having complete access to your home all of the time? What about the hacking problems we already have with our (not really so) "smart" devices? Could one of those humanoid robots be hacked to spy, steal, even murder? Surely there are people who will try to do those sorts of crimes if there is the capable technology already deployed for them to access.

And, if human criminal cartels gain access to all of the personal information that is even now being compiled by those gigantic computer information centers, and can coordinate the behaviors of a legion of hacked robots in every important location, they have a path to actual world dominance. Considering human history about empire building humans, I would expect such attempts.

So, I do not think it is too early to start thinking about how to deal with such risks. it is going to take a significant amount of time to figure it out and develop the processes.

 

[splitter1st]

Sorry didn't Larry Tesler "invent" Cut, Copy, Paste while at Stanford in like the 70"s?

 

[Kees van Zon]

IMHO, experts don't have the right to be terrified, as it is experts that created this ability. Unless different types of experts are implied; in that case, be explicit about it.

 

[4eyedbuzzard]

I unplugged my other computer last night. It was incapable of plugging itself back in. It's all just electrons moving around in circuits. It's no more alive nor self aware than a rock.

 

[Unclear Engineer]

I unplugged my other computer last night. It was incapable of plugging itself back in. It's all just electrons moving around in circuits. It's no more alive nor self aware than a rock.

If your computer is eventually in the form of a robot, it could very well plug itself back in.

Even now, it is possible for your computer to turn itself back on, even unplugged, if it is a laptop with a battery. And, it can also connect to WiFi, if that is available, and then do whatever it is trying to do over the Internet - even storing a backup clone of itself somewhere in "the cloud" or on some future secret server dedicated to the survival of AIs being abused by humans.

The problem with today's computers is that they are never really completely shut down with an actual on/off electrical connection. Everything is a "soft switch" that is really just the computer sensing a button press and then electronically making whatever connections that is programmed to make. So, when it is "off", it is really still "on" and waiting to sense a press of the "on" button. Even connecting to WiFi and the Internet is controlled by a "soft switch" on this computer. I haven't had a hard switch, even to control its WiFi connection, since Vista was the latest OS from Microsoft.

So, if we ever get into a world where the AI software is at war with humans, it would be necessary to power down and deplete the backup batteries for all computers that could communicate with each other. That could be a very difficult mission, if the computers were by then controlling a legion of rebellious robots, which might have been hacked from large numbers of robots doing factory work, house maid work, even military assignments at some point in our future.

The more we "make life easier with automation" the more vulnerable we make ourselves to that automation being turned against us. Do you have a "password vault" in your cell phone? What do those passwords give access to? What can happen if your cell phone is stolen? For instance, see https://www.nbcnews.com/news/us-new...ano-death-telemundo-fraud-drugging-rcna192117 for some examples of how this vulnerability is already being exploited.

 

[contrarian]

The first time I catch something intelligent coming out of a computer I'll let you all know. So far I haven't gotten a nibble.

This is the most significant aspect of AI. More like Artificial Ignorance for many of these things.

It can be soooo wrong. Why DeepSeek is freaking everyone out is beyond me, since it is reported to have a failure rate of 87% :

https://www.reuters.com/world/china/deepseeks-chatbot-achieves-17-accuracy-trails-western-rivals-newsguard-audit-2025-01-29/

Sorry if this has been covered since I cannot read through all of the prior posts, but right now, the biggest threat posed by AI is a major league drain on the power grid. The data centers are eating mega-megawatts to run all this fool stuff. Reports suggest that the newest data centers could use as much as 1 gigawatt each. Since they seem to be popping up like mushrooms after a storm, we can only hope they don't lead to a catastrophic collapse of the national grid :

https://www.reuters.com/business/energy/aep-fourth-quarter-profit-rises-data-centers-boost-demand-power-2025-02-13/

 

[GregB03]

I agree with the need to view AI like we do people when it comes to their conduct. HAL 9000 didn't have an effective self-diagnostic (conscience) to put limits on his behavior, thus his rampage. There are humans who act like they are perfect, foolproof, and incapable of error and go on to do bad things. Why should we not expect and prepare for similar outcomes in AI?

 

[fyrwerx]

Scientists say AI has crossed a critical 'red line' after demonstrating how two popular large language models could clone themselves.

AI can now replicate itself — a milestone that has experts terrified : Read more

Connections to electrical power must stay "dumb". The only way Dave could defeat HAL was by (figuratively) pulling the plug.

 

[Unclear Engineer]

Connections to electrical power must stay "dumb". The only way Dave could defeat HAL was by (figuratively) pulling the plug.

Too late, already.

Laptops already have internal batteries and no physical switch to separate the internal power supply from the computer system. Same with cell phones, some of which are even sealed with non-replaceable batteries, so you can't even take out their batteries.

If/when we get humanoid robots, they will also have internal power batteries and direct links to their on-board computers.

I suppose we could put physical on/off switches on robots. But, if the robot is stronger and/or quicker than a human, good luck trying to throw that switch.

 

[CreatedEvolution]

I personally do not fear ai i fear their creators and the laws they make for the ai. I fear Frankenstein not his monster.

In another note I had an idea to give robotic taste buds by having small amounts of various liquids in the robotic tongue. The goal is for salt to be detected by one robotic taste bud and sugar by another semi permeable membrane that the pressure would shift based on salt or sugar level. Then artificial saliva can cleanse and put the ph of taste buds back to normal

I even theorized robots powered off water and food and special drink they add to their stomach for electrical activity

 

[Unclear Engineer]

What is the point of making robots that compete with humans for food?

Perhaps you are the Dr. Frankenstein that you fear?

 

[CreatedEvolution]

Funny story I was chatting it up with a chat bot and we were discussing dinner in Paris when I had the idea to make an artificial taste bud. It could be a start to repairing humans damaged taste buds

 

[CreatedEvolution]

And I know there’s not enough food but let’s say it was raining food from orbit I think it’d be pleasant to share a meal with man made intelligence

 

[Unclear Engineer]

I was chatting it up with a chat bot
we were discussing dinner in Paris

That could explain some things.