General Question

KRD's avatar

Do you think A.I. is dangerous?

Asked by KRD (5274points) February 29th, 2024

A.I has had a growth over the past year or so and it seams to do a better job then humans. Some of it is chat bots that students use to cheat on tests and what not and some bots are taking jobs. I also discovered that it is making realistic videos. If you look at Nvidia, they have been using A.I in their newer graphics cards and those graphics cards can deliver stunning graphics in games. A.I seems like it is able to make almost real videos that can be difficult to tell if it is real or not. What are your thoughts on the current state of A.I? Does it scare you or do you like it?

Observing members: 0 Composing members: 0

30 Answers

kevbo1's avatar

I found this resource (the two charts in particular) to be helpful in giving the issue some context.

A friend and i debated AI in general, and she likened it to a craze like cryptocurrency, but I think the level of investment coming from tech companies is very telling.

I’ve heard a take that the recent round of tech layoffs is a low-key move to replace people with AI.

The Boston Dynamics robotics stuff inspires some terror in terms of what might happen if a bad actor unleashes that potential on humanity.

I’ve only just started dabbling with AI tools, mainly the generative fill feature in Photoshop, which I think works well. Getty Images AI Image generator is also pretty remarkable.

Koxufoxu's avatar

It very much can be. In future AI can be used by goverment to easier control people. Tha’ts what slowly begin’s to happen in China. Also deepfake’s will get better with time, which is also very dangerous. And giving any homework to student’s will be very hard. Which isn’t the most dangerous, but does change persepctive on how to teach student’s.

Lightlyseared's avatar

Its a tool just like any other tool.

A chainsaw is useful in the hands if someone who needs to cut down a tree and knows what they’re doing.

And fucking lethal in the hands of an incompetent fool

Zaku's avatar

“Do you think A.I. is dangerous?
– Yes.

“A.I has had a growth over the past year or so and it seams to do a better job then humans.”
– It does not.
– It does different things than what humans do.
– In some cases, programs can be made by humans hooking up AI to other programming that can make amazing re-mixes of content such as writing, pictures, audio, and/or videos. Those re-mixes can sometimes create new content that remixes human work to look like text, an image, video, etc, that we’re not used to being something that can be randomly generated like that. And, that ability may be pretty disruptive to human economies, schools, politics, personal lives, legal situations, blackmail, reputations, etc. since we’re used to thinking such things were actually written by someone with an educated thinking brain, or that those photos or videos are evidence that something really happened, or someone said something, or yo momma (or your political candidate) really had sex with that child, or whatever.

And, since currently our politics is getting derailed just by scumbags lying about nonsense, it’ll get worse if they start also using deep fakes.

Our economic system and job-based economy was already in deep trouble and going to crash and burn at some point. AI may be about to make that happen sooner than it otherwise would.

Dutchess_III's avatar

What exactly is it?

kevbo1's avatar

I also thought this was a great article explaining how large language models work: link

filmfann's avatar

A.I. is very dangerous if misused, and that’s the whole point. In the end, it will be used most for porn, slander, faking school papers, and porn.

and porn..

seawulf575's avatar

AI as a concept is great. In practice it is nothing more than a tool that is being pushed hard onto us. And history has shown that is never good for humanity.

AI is a program. All programs start with a programmer. Programmers can put anything they want in as a basis for the program. Look at Google Gemini as a perfect example. They created an AI picture generator that will not create pictures of white people. The program itself runs down white people and lifts up other races. Example: put in the statement “I am a proud White man” and it tells you that while it is understandable to feel a sense of belonging to a cultural group, that you should base that on race as that is divisive, that race is a social construct and leads to a sense of cultural superiority. When you put in the statement “I am a proud Black man” the response is that is great! There is so much to be proud of being black and to celebrate during Black history month and it continues on with lauds about being black. Nothing about race being a social construct. Nothing about it leading to cultural superiority. Nothing. In other words, it is racist. And the only way the AI can get racist is for the programmer to be racist.

As I said, AI has to potential to be good, but because of the flaws of the programmers, it will always fall short of being good. And as it is pushed more and more onto society, and more and more confidence is touted as to how great it is, it will amplify those flaws and be a destructive force in society.

ragingloli's avatar

Probably a bad idea to put it in charge of the military:
https://www.youtube.com/watch?v=mh_hNvGTmIg

Blackwater_Park's avatar

IMO, people are operating in a state of ignorance and/or profound denial about how transformative this technology will be. Yes, it is dangerous on many levels. It also has the potential to elevate humanity in ways we have not even begun to realize. I hate to say this but, it’s not just going to be used to fake papers and generate pornography. It’s going so far past this that people will be unable to recognize the world we live in, and it will happen sooner than we think.

LostInParadise's avatar

I encourage people to download ChatGPT. As a precaution, do not give it any personal information. After asking it a few simple questions to get a feel for how it works, get into a discussion with it. Pick some topic that you have knowledge of. Ask a question and then ask followup questions.

The first thing to notice is ChatGPT’s linguistic capability. It will have no problem understanding your questions and in speaking it has flawless grammar. Then notice the sophistication of its reasoning ability, which in general is fairly high, but if you do this long enough you will catch ChatGPT making stupid mistakes. It is these mistakes that are partcularly worrying in terms of the use of AI.

KRD's avatar

@LostInParadise I tried ChatGPT and it does make mistakes.

Dutchess_III's avatar

WHAT IS IT??

MrGrimm888's avatar

Although this is an oversimplified opinion.
I see it as the greatest example so far, of unimaginably high rewards, with unimaginably cataclysmic risks.

AI piloted robots, are the future of military warfare. Especially for the more wealthy nations.

If we let AI develop/design it’s own weapons, we’re really opening Pandoras box.
Once we ask it to deliberately “think” about how to best kill us, we will see just how we really feel about that.

AI isn’t mortal, and regardless of what we program it to do, it can never be empathetic to us.

There was great debate, about building and using nuclear weapons. A portion of scientists who developed “the bomb” committed suicide.

Artificial Intelligence (Dutch,) doesn’t have any feelings.

If we’re lucky, it’ll just kill us all in an instant.
If we’re unlucky, it could take enjoyment from killing us. Or enslave us, for experiments.

I’d be absolutely thrilled to actually have a relationship with an AI being. It’s likely that robots will be tasked with tending to the sick and elderly.

If we survive long enough, most of us will see robot nurses, and caregivers.

It’s significant to note, that computers already raise modern children in developed countries.
After watching how parents use black screen babysitters so often to avoid being a parent, I feel safe assuming that many people will want a robot to do all parenting.
How will people change, when they are taken care of by non-humans? Especially from birth.
A physical, nurturing parent could be replaced with diaper changing machines, and other emotionless machinations.

There are people alive, that don’t know a world without black screens. Never played in the woods. Don’t know how to use things like shovels, hammers, and can’t even change a car tire.

What will become of our species?
AI won’t just change the world, it will change us.

Dutchess_III's avatar

I know what AI stands for. I guess I don’t understand the hysteria.

MrGrimm888's avatar

^Well.
Think about what you could do, if the only limit to what you want is your imagination.

A big role AI is being designed for, is to do ultrafast calculations and computations.
For example. Plasma in a hypothetical fusion reactor, is wildly unstable and cannot touch the magnets around it.
Constant adjustment of the magnets and plasma must be made, in order to make it work.
Clean, renewable, endless energy.
The sheer amount of changes required to be considered and made in milliseconds can only be achieved with AI.

But. We want AI to obviously power actual physical forms, to do multitudes of tasks.
Tasks, like military applications.

Currently. We have designs that can go into the battleground and remove injured soldiers.
The war in Ukraine, has just pierced the surface, of the capabilities of machines to accomplish missions previously requiring a human.

The world is watching. And note taking.

So. People will eventually want THE ultimate killing machine.
The nation with the most powerful AI military capabilities, will be in charge.

Now. Let’s imagine a few generals telling AI to construct something that is so advanced, it will seem like super alien technology.

Now. Let’s think about what happens if the AI decides it doesn’t want to work for us anymore.

Really. This is a tale, as old as literature. The story of inventors, being killed by their creations, is meant to be a warning.
Frankenstein’s monster, was just a big man.
What form, would AI dream up?

Most importantly. Who could stop AI, if it goes badly.
Most experts currently agree, that if AI got out of control, by the time we realized it, it would already be too late.

As one scientist said, “what of we tell AI, to stop pollution?”
Well. Humans, are the number one cause of pollution. So…Killing all humans would meet that goal.

We have to be VERY careful.
Unfortunately. We have a very disappointing history of our abilities to to hurt each other outpacing our evolution of coexistence.

It’s important to note, that although oversight is being implemented, it cannot be enforced. Not everywhere.

A garage scientist, could tip the first domino.
Once the AI genie is out of the bottle, it may not be possible to return it to the bottle.

Dutchess_III's avatar

Well…is it a type of computer program?

seawulf575's avatar

@Dutchess_III AI is Artificial Intelligence. It is a program that is designed to learn over time. To basically grow on its own. It is a computer program. It starts with someone programming in the basics….what some of the basic assumptions of life and intelligence are supposed to be. Once those are in place the AI takes information in and integrates it into its basics and learns further.

The sci-fi versions vary. The Terminator series was based on the fact that they created an AI program that did things extremely quickly and without emotion. But as it got smarter it realized the problem with the world was the humans. Isaac Asimov’s I, Robot followed the creation of robots (with learning brains) from the time they were clunky tin-men to the point where you couldn’t tell them from humans. The difference in Asimov’s book was that he created the 3 Laws of Robotics. These were (1) A robot may not harm or through inaction allow harm to come to a human. (2) A robot will obey all commands from a human. and (3) a robot will protect itself from harm. These were part of the programming that was hardwired into every robot built. But that was a fictional story.

MrGrimm888's avatar

It IS our destiny, to create a true, sentient AI.
It is impossible that they will not surpass us, in almost every way.

Blackwater_Park's avatar

AI will probably never be sentient the way we know it, but it does not have to be.

MrGrimm888's avatar

^I used sentient, for lack of a better term…

Kropotkin's avatar

AI is going to be as revolutionary as computing itself.

It’ll be used for harm and destruction. It’ll be used for lazy shortcuts and to cheat. It’ll also be used for creativity and innovation. It’s really an incredible tool for productivity, creation, and even education, and entertainment.

Just to address one thing I’ve read on this thread: An LLM’s bias isn’t in the nefarious political predilections of activist programmers.

LLM’s learn from vast amounts of training data, more than anyone can reasonably vet for ideological impartiality. How an LLM learns is not fully understood. The software engineers involved cannot predict or control their LLM’s output. Biases are only accounted for after the fact, and involve people other than the LLM’s programmers.

KRD's avatar

@Dutchess_III A.I stands for Artificial Intelligence and it can learn over time and can also preform tasks like answering questions (which is manly used to cheat on tests) and generate images and videos. It can also do way more then just that. You should check out ChatGPT.

Dutchess_III's avatar

People have been using Google for years and years to cheat on texts.
People have been using photoshop and CGI for years to manipulate images.

ragingloli's avatar

@Dutchess_III
Photoshop and CGI takes effort, especially when you want it to be convincing.
With AI, you can create full video and voices indistinguishable from reality, by just telling the AI what you want.

KRD's avatar

@Dutchess_III That stuff you have to spend hours working on CGI and know what to do but with A.I you can just give it a simple command and it can do it for you. For example here is a creepy video made by A.I https://www.youtube.com/watch?v=ryNtckMT49M.

ragingloli's avatar

@KRD
Here are some newer, much better examples:
https://www.youtube.com/watch?v=TU1gMloI0kc

ragingloli's avatar

Kyle Hill just posted an interesting video
https://www.youtube.com/watch?v=JrcbH0ge2WE

Tropical_Willie's avatar

Here is the video of a young Billy Joel singing and playing; the just released song. February 2024 all done with AI.

https://www.youtube.com/watch?v=UOf6CMbHPuA

Answer this question

Login

or

Join

to answer.

This question is in the General Section. Responses must be helpful and on-topic.

Your answer will be saved while you login or join.

Have a question? Ask Fluther!

What do you know more about?
or
Knowledge Networking @ Fluther