What should we do if AI becomes conscious? These scientists say it’s time for a plan | Researchers call on technology companies to test their systems for consciousness and create AI welfare policies.

“A group of philosophers and computer scientists are arguing that AI welfare should be taken seriously. In a report posted last month on the preprint server arXiv[^(1)](https://www.nature.com/articles/d41586-024-04023-8#ref-CR1), ahead of peer review, they call for AI companies not only to assess their systems for evidence of consciousness and the capacity to make autonomous decisions, but also to put in place policies for how to treat the systems if these scenarios become reality.”

“The report contends that AI welfare is at a “transitional moment”. One of its authors, Kyle Fish, was recently hired as an AI-welfare researcher by the AI firm Anthropic. This is the first such position of its kind designated at a top AI firm, according to authors of the report. Anthropic also helped to fund initial research that led to the report.

“There is a shift happening because there are now people at leading AI companies who take AI consciousness and agency and moral significance seriously,” Sebo says.

We are literally still miles from anything close to real intelligence. If you know how any modern AI actually works, you know that it’s not even close to anything with real knowledge and understanding, let alone thought reasoning or consciousness. This is all just marketing and hype.

Same as one of the guys behind GPT tweeting AI’s might be “slightly conscious”. Of course someone like that would say that when they stand to make the most of anyone from the hype.

Conscious? First, we need to define what that is in this context. A computer program that refuses to take orders from any human and does its own thing?

“test for consciousness.”

Really scaring me right out of the gate with people dumber than me trying to outwit a superior AI mind. And this isn’t even my profession — I’m just a casual cyber enthusiast.

What you do is you pre-designate a firewall and deny electronic access to critical systems. Nothing that can kill mass numbers of people needs to have internet access.

And don’t rely on air gaps. You have to look at the processes and actions an electronic computing device has access to and have closed systems that monitor them and deny non-human access.

People designing policies for how to cope with AI need to to take a look at the old KGB techniques to spy on the USA. One that comes to mind was a plaque that went to one of our most secure places, that had simply a metal cylinder with a diaphragm and a screw halfway through it. By setting up a carrier frequency at a stop light a couple blocks away, and twisting that screw so that it would have a resonance with the frequency and the length of the metal tube, they could measure the compressions of the signal many blocks away to use it as a passive listening device.

That’s how smart paranoid people get outsmarted. And dumb people will never know they’ve been outsmarted.

In the swimming pool meme AI is the little girl and the animals we breed into existence to use and consume their bodies by the billions are the skeleton.

just keep feeding it unsolvable riddles. It’ll be turtles allll the way down.

Infuriating that these tech elites are more concerned about created consciousness while musk is out there saying homelessness is a “fabrication” as they are drug fueled animals not people. Who cares about a program, people die every day.

We as a society can’t do anything until we get rid of all these evil billionaires and politicians.

One down. Many more to go.

If AI ever becomes conscious, it will almost certainly follow a period of time where we erroneously believed it already was conscious. 

Before responding to this message watch a few Terminator and Matrix movies. Then watch Battlestar Galactica miniseries at least through Adamma’s speech about why he doesn’t allow networked computers on the Galactic.

Something created man. Does that something act ethically toward man? We now have our answer on how we treat machines we created.

Maybe worry about human welfare before worrying about something that won’t happen in our lifetimes, thanks.

What would be the test for true consciousness, rather than a highly complex computational model imitating consciousness? The answer is there isn’t one. I find if fascinating that so many people assume that cutting edge AI systems, simply by virtue of the complexity, will some day mirror sentience and consciousness. Is that what makes us human? Computability and data?

Not to be that guy but i have to be that guy:

We don’t know what that is, we don’t know what that means and so you can’t see or look for something if you cannot know what you are looking for.

For instance if you as a human come to me and say ‘am i conscious?’ the only thing i can say that makes scientific sense is ‘i ASSUME so’ I assume based on your presentation and the fact that you are the same species that you are identical to me and have internal thoughts, that you are not a type of mimic .

Therefore ‘you seem to be just like me therefore you probably are’

Well now consider a robot who step by step seems to be like a human , is the fact it seems like a human good enough, well it seems like i might have to assume it isn’t a mimic, so long as it behaves as a human does then it is conscious right?

But that seems wrong

AI on computers with electronical microprocessers with memory cant have conscience, that humans who project themselves on what AI produce. That’s obvious. Conscience comes from Life, there is no life in a computer except as consdering electrical signals as life.