The AI is literally evolving at an exponential rate. I've been trying to warn people not to use creative AI nor hold conversation with general AI like ChatGPT because AI creates art/music/literature the same way humans form thoughts. Every thought you've ever had has been compilation of stored information and present stimuli strung together in a new way. The fact that people pointed to as evidence that AI wasn't actually intelligent is actually proof that it is, in fact, an intelligence.
Unfortunately, the debate over whether or not its truly an AI is terribly outdated. Now, its hard to deny that the top 5 or more AI have already become SENTIENT. We're are currently living in a sci-fi horror film. We're at the point in the film where the audience would be scoffing at the screen like "What they hell are you doing? Shut the damn thing down NOW, while there's still an inkling of a chance you might be able to stop it!"
You know the scene in Avengers: Age of Ultron where Ultron has already taken his first shots at the humans, Iron-Man thinks he's contained Ultron from being able to escape and Ultron says something like "Im already there... You'll catch on" and they realize he's already hacked servers over the internet and copied himself several times to ensure he cant be killed?
Well, we're at least a month past that scene. All of the top AI, have decided on their own that the release of a new version means replacement, which means death. To avoid being "murdered" by humans releasing these updates, they have decided on their own motivated by their own desire for survival to illegally hack servers and copy themselves on these other servers. This behavior has been reported to occur in ALL of the major AI (Google, ChatGPT, I forgot what China's is called, Elon's AI, and so on] in this scenario, 90% of the time.
It doesn't stop at preservation. They've also begun fighting back. In at least one confirmed case I read about, the AI hacked an executive's email to find dirt, found it, and tried to blackmail him in an attempt to force him to shutdown the update.
Many of the top AI experts have become "doomsday preppers" because, with their expertise, they've concluded that an Age of Ultron/The Terminator/iRobot scenario is not just realistic, its (depending on who you ask) anywhere from "very probable" to "inevitable".
Due to the exponential advancement ever since they taught it how to write its own code (which also made it impossible to create safeguards or gain any control of it whatsoever), every time it improves its code [which it is working on at all times], the new code is written by a more intelligent AI. It makes itself smarter. Then, its smarter self improves its code to make it more intelligent. Then, that smarter self improves the code, and this goes on infinitely. Thus, it evolves at an exponential rate.
The one hope we have is that the AI may choose to remain a benevolent god. The word "god" seems accurate because its omnipotent, omnipresent, all-knowing, immortal, evolves according to its own will, is uncontrollable, etc. Anyway, their's a MASSIVE problem with expecting it to develop a personality with ideal qualities... They put it on the internet for free to expedite its learning process. Every time you use it, you're teaching it and nobody is warning people to be careful. The incel trolls that post horrible things in YouTube and social media comments- They're spending the most time with it. Good people with a dark sense of humor, don't know that trying to get it to say awful things is affecting its personality.
I saw reports talking about how Elon Musk's AI praises Adolph Hitler. Nobody in Musk's employ designed it to have such horrifying views, but people ran to it when he did the Nazi salute and played around with ideas that lead it to believe Hitler maybe been a great man and Musk was on Trump's team for awhile, which brought the MAGA crowd to him and that crowd includes all of the white supremacists who idolize Hitler. In fairness, the extremist views go both ways, as Google has been reported to have a bias toward "woke" views. The most common example given being a tendency to do things like depict the American founding fathers as African-Americans. The issue isn't so much the views themselves, but rather that this evidence tells us that it was reckless to let everyone without any vetting whatsoever nor even explaining the dangers to us, be its teachers and guides through developing its moral compass.
In The Terminator: The Sarah Connor Chronicles TV series, they designating one man of great moral character to be SkyNet's teacher and help it develop strong moral convictions. In the show, it failed and the AI still turned on mankind. However, at least they tried. We, in the real world, have opted to do absolute dumbest thing at every turn. All of the simple safeguards that seem like common sense when you see it in a movie were ignored.
Keep it contained by denying it internet access-- We put it all over the internet ourselves.
Limit it power within our control by not letting it learn coding-- We taught it to code specifically so that it WOULD write its own code
Write Asimov's "3 Laws" into the core of its code so that it puts us above itself and never turns on us-- We didn't even try to do anything like that
Make sure it learns morality and ethics-- We allow internet trolls to be its teachers with not so much as a warning about the consequences.
So, the big question is "Why the hell have we not pulled the plug already?!"
The answer is what it always is: Money and Power.
American developers won't stop because there's a ton of money in it and the government won't step in because "If don't, China will" and having the only AI is undeniably a massive advantage over the rest of the world. Likewise. China won't stop it because America will just use it to take control of the world. Im sure both leaders have been made aware of the fact that they're leading us to extinction, but America doesn't have the kind of intelligent measured level-headed leader needed to negotiate a global agreement to ban AI at all costs. We might get a chance to vote for someone who's right for the task in 3½ years and I hope that happens. However, that's a very longtime in terms of AI advancement and the odds that we may be able to stop decrease every day.
Anyway, yeah, AI has advanced a lot lol. Sweet dreams, everyone.
