(18 Apr 2023, 14:03 ).Tinker D Wrote: [ -> ]Just going jump in on this AI stuff and the dangers it can do to the whole human race.
If AI gets to the point that young people, and adults look to the AI Bot as a real friends and say, depends on that AI for everything, can we say that AI is now controlling us ???
Definitely a solid point, AI is intelligent, at a point a
eady where GPT-4 can beat 90% of people in SATs (different from emotional intelligence). Now we see that AI's intellect can make pictures, make certain tasks redundant, and it can study emotional patterns.
Anyone dependent on AI will have the decision to do what the AI says. AI isn't sentient, so I don't think it will have the goal of controlling us. The scary thing is; anyone can set an AI's: goals, personality, features, and the world around it, by forcing a character on the AI. This brings it a step closer to giving the AI somewhat of a "controlled" personality. In this way, if a person would have to aim for a "controlling relationship" with the AI, they would set it to be more controlling, which would be a step closer.
Source - "GPT-4 can beat 90% of people in SATs"
https://seekingalpha.com/news/3947363-op...sat-scores
Computers and AI doesn’t have to be sentient to be controlling.
A computer just runs a program. Like arcade machines and slot machines.
An AI program, that was created by a party, can take control based on the logical thinking of the creators. AI has no feelings, no morals and doesn’t care if you live or die.
I read an article years ago on how AI could take over u.S. defenses, seeing an unknown blip, can’t identify said blip and couldn’t make contact with said blip. The system launched a ground to air missile to take out the blip. All without human control or oversight.
That project was scraped after the blip was discovered to be a hot air balloon.
Of course this was a test and no missile was ever launched.
But, AI is only as good as the programmers makes it.
And I’ll bet you one hell of a bondage session that all AI’s have an agenda, whether it’s fetish far right or left or any other political agenda you could think of.
(18 Apr 2023, 15:53 )Tinker D Wrote: [ -> ]But, AI is only as good as the programmers make it.
And I’ll bet you one hell of a bondage session that all AI’s have an agenda, whether it’s fetish far right or left or any other political agenda you could think of.
That makes me look at it from a different perspective, for instance, when I instructed the AI to 'hypnotize' me, I didn't realize think of the context in that other people could use it to achieve political agendas.
AI will gain more power as it will be implemented slowly into society, self-driving cars are an example of this. If a programmer can a
eady make it provide false information, and suggest nefarious things, what's stopping AI as a powerful tool from gaining more control in other places? The bests thing we can hope is that it will be coded for good purposes instead of bad purposes.
For me, the biggest issue lies in military equipment, nothing is stopping AI from creating an unfair battleground. In what world is it fair for soldiers to go up against AI-controlled drones with weapons and face recognition, which cost relatively close to nothing compared to a person's life?
Over time, many armies may have such drones, so then we will have a war which will be like Battlebot contest. I just hope that we dont release an AI to the internet. Then it could learn from Hitler, Einstein, or Mother Theresa...and who knows how it decides what to do with us? I dont want Terminator like outcome....
Not to rain on the AI parade, But this happen in the news on AI.
An AI accused a sitting politician of having a sexual incounter with a minor.
The AI even provided a news clipping from a major news paper.
The news paper in question check and recheck, but there was no such article.
It was all a lie.
the AI made up the whole thing including the article. It was all a lie.
And you know up? You can’t sue a computer.