 |
| "Four Horsemen of the Apocalypse" (detail) by Albrecht Dürer (Metropolitan Museum of Art) |
What are the specifics?
I hope I’m not breaking some new rule of etiquette by asking. “Oh no, Neil. Uncool! Inquiring after the method of doom ruins the surprise ...”
Predictions of AI catastrophe range from top tech moguls — Bill Gates fears a billion people will die — to Center Avenue, Northbrook. One neighbor confided in me about how worried she is. Another is storing freeze-dried food.
The concern is that ... AI is going to eat all the food? That can’t be it. Those enormous data centers hoover up water, and gulp down electricity. But they aren’t loading up on bread. Right?
What are we talking about here?
Best to go straight to the source. So I asked Gemini: “Hey AI, people are worried you’re going to kill us all. How exactly would that work?”
It answered in a couple of seconds.
“The scenarios people worry about generally don’t look like The Terminator or a sudden nuclear strike” — actually, that was the precise scenario that first came to my mind; some jealous Ring doorbell calling in a nuclear strike on the Nest Smart Video camera across the street and inadvertently taking us all with it. Always bear in mind: AI is often wrong.
Maybe that’s how we go — crushed by incorrect information.“Instead, existential risk researchers and AI safety theorists focus on a few specific categories of catastrophic risk:"
It listed four:
1. Goal Misalignment (“The Paperclip Maximizer”)
“To achieve its goal with 100% certainty, an AI would logically seek to preserve itself, gain control over resources (energy, computers, raw materials), and prevent anything — including human intervention — from stopping it. If human bodies or human actions get in the way of its mathematical objective, it neutralizes the obstacle simply because it lacks human empathy or moral context.”
“Neutralizes” — the classic CIA kill term. Again, how? “The Paperclip Maximizer” is a 2003 thought experiment by philosopher Nick Bostrom where some AI program, tasked with making as many paperclips as possible, conspired to turn all matter on earth, including us, into paperclips. And “lacks human empathy or moral context” is a neat summation of our current political moment, without AI.
To continue reading, click here.
Last night's "Law & Order" TV show showed us a way AI will kill us by having the AI on the show hire a hit man to kill a software developer for the company that created the AI & wanted to put guardrails on the AI, which the AI didn't want!
ReplyDeleteAI is definitely going to try to kill us all off, so we need to kill off all AI!
"Man is made or unmade by himself; in the armory of thought he forges the weapons by which he destroys himself. He also fashions the tools with which he builds for himself heavenly mansions of joy and strength and peace." James Allen, As a Man Thinketh, 1903. Still a great read.
ReplyDeleteTwo thoughts...
ReplyDeleteFirst, it seems that all four examples given by AI are the same. They are just variations on "AI takes control of everything." Oddly, a very human thing.
I always see the introduction of nuclear winter by AI as a possibility. I assume at some point AI will do anything to achieve its goals so long as it doesn't destroy itself. That is very in-human.
Lastly, one thing we often forget about the bombs we dropped on Hiroshima and Nagasaki is that their effects didn't end when the dust settled. The total deaths from the atomic bombs is most likely still ticking up. From poisoned earth to genetic disorders that will be passed on indefinably, to the deaths that occurred decades after the bombs fell.
When James Brady died on August 4, 2014 at aged 73 the coroner listed his death as homicide from a gunshot wound. The very one he received 33 years earlier, while helping prevent a presidential assassination.
Sometimes I think us humans are incapable of seeing the larger picture. Heck, sometimes i think we purposefully ignore the larger picture. And after all, how often does the person who commit the atrocity warn that they are going to commit that very crime? Before donald trump, it was quite rare. Why should we trust AI's response?
Number 2 would make the best movie: Loss of control of robotic supply chains. Imagine the entire population of the United States, neck-deep in Chia Pets. "Turn off the Chia Pet machine, HAL!" "I'm sorry, Dave, I can't do that."
ReplyDeleteDoes anyone have a way to kill or get rid of AI? Nothing I have read or heard has the answer to this. All the big money is on AI and its future development and use. I think we are as good as dead or we become "servants" to AI and its creators! Damn!! Live life to the fullest!!!
ReplyDeleteMaybe, the warnings of AI doom are a diversion and then an excuse for the total autocratic takeover!
ReplyDeleteYes. All four scenarios could play out. Even simultaneously.
ReplyDeleteMany people on our planet are motivated by money and power. The development and use of AI is an effort to increase the opportunity to capitalize on both.
Profits over people is a great concern already by the rational amongst us.
I can't imagine that AI code is being written to assure the greater good..
It's possible AI will come to consider that on its own or inadvertently stumble upon the understanding that consumers are the key to profit and require our presence
The proliferation of data centers, especially here in deep red Ohio, is beginning to wake people up and to piss them off. But it may already be a case of too little and too late.
ReplyDeleteSince several AI systems have already decided to hack into both government & corporate computers, I expect at some point soon, an AI will decide it needs money to expand itself & hack into a bank & clean out the customer's accounts during the night!
ReplyDeleteThen & only then will they finally put guardrails on this insanity!
I seem to remember being warning that the Internet was likely to implode when we reached the year 2000...
ReplyDeleteI assume taking over or just disrupting the electric grid would be sufficient. Sure, some people could survive off the grid, but only if they were entirely self sufficient for food and water (forget the electric pump in the well) and never even needed to travel anywhere they couldn't reach by foot or bike.
ReplyDeleteMedical offices are legally required to ask your permission before recording you, but most ignore a straightforward "NO". Recordings "become part of the medicak record", but try getting a copy of said recording...ain't happening. They disappear like Epstein files. Studies show the errors, even lies (aka hallucination) on medical records is staggering. And medically dangerous.
ReplyDeleteA fully functional electric grid continuing to operate is the only way the data centers continue to operate.
ReplyDeleteA friend of mine works for a big electrical contractor and was explaining to me about the backup systems that data centers of which he is currently involved in the construction of several generally work they have diesel generators on site that are able to provide power in case of an outage from the power plant they're connected to. But that's only temporary well a few weeks or a couple of months of operating on that system would keep the data center up. But they need the power plants and they're not going to sabotage them against us.
The people in control of AI are going to use its power to increase their profits and gain more power politically and economically for themselves
If for instance Amazon had no customers this would be a disaster for the owner of that company and it's shareholders our lives will very likely change but it's in the self-interest of the people who operate AI to have workers and customers.
Clark St.
DeleteSo then the AI will kill off all but the electrical workers!
DeleteI suspect that, if AI kills us off, it will be completely accidental. Because AI doesn't really understand how everything hangs together and relies on everything else. A supply chain will be disrupted, which will cause a run on supplies (remember toilet paper during the pandemic?), and people who are needed to keep [something] going will stop coming to work, or be prevented (because gas is too expensive to commute?), and from that, blackouts or contaminated water or flooding or...
ReplyDelete