There has been a lot going on when it comes to AI over the past few months. First it was the controversy with AI art. Then it became another big deal when the writers strike happened and people thought movies would be written with AI. Then a few weeks ago the entire opening for Secret Invasion was done by AI. I didn't care too much until recently with the stuff revolving Erica Lindbeck since that's in my industry. What are your thoughts?

https://www.thenews.com.pk/latest/1088772-dcs-superhero-slump-deepens-with-the-the-flash

In my opinion I think it is bad.

Scammers are already using AI to copy a real person voice. Making others that know that person think they are in trouble. In My state a scammer used AI to imitate a person's voice. They were able to hack and get this person contact. They used the copy voice to get the person's family to wire some money.

Now, this might be far fetched. But what if the Govt started using Ai in a bad way to go after citizens. I could see the US government doing that. This is the same govt that tried to cover the Gulf of Token being staged and Waco incident up. This is the same govt that Spied on us via the Patriot Acts. It has come out that they even tried to censor groups on social media. It is possible the US Govt could use AI in a very abusive way of power.

Finally, could AI turn against us. That may seem far fetched but it may be possible in decades from now. It sounds like a science fiction movie. But what if the AI robots and computers turned against us. It could very well happen.

These are my reasons that I think AI is bad in general.

It might be a little hypocritical of me to say that AI is bad considering that I have used Chat GPT more times that not, but it can be used maliciously by not well-to-do people on the internet.

The biggest problem I have seen constantly while being on social media was the unauthorized usage of others art to create AI art pieces. That always comes off as scummy to me. You have artist who put time and consideration into making nice pieces of fanart, original art pieces, backgrounds and scenery that all have something unique to them. For that to just be yoinked and dumped into somebody's AI art prompt with no second thought to how this could effect the artist they steal from is saddening.

tornadocam touched on this as well and I think that Erica Lindbeck/ Futaba AI controversy also hits on this is the inappropriate usage of AI to mimic voices. The growth of AI is crazy to see. Two years ago, if you would have told me that there was an AI cover of a Singer covering some other singer's song I would not have given it a second thought. It would have been rough, jagged, and a bit robotic. Now, my feed on Youtube would have some stuff like Kurt Cobain covering Buddy Holly or Boulevard Of Broken Dreams with little to no resemblance to it ever sounding like what it might have sounded like two years ago. If your using it to see what it sounds like then go ahead. But if your using it to spread hate, misinformation, or just using their voices to make a quick buck, then your just being an ass.

I read somewhere that AI useage was one of the causes of the current ongoing writers strike....something needs to be done about this

I actually had an epiphany while I was typing a post for the reality tv post, why can't we just have AI do reality tv instead of doing the writing and hologram acting, I mean watching AI destroy themselves this way could be entertaining

I mean with AI becoming more of an issue with Hollywood ( actors wanting AI protection) I feel like this might cause more harm than good without setting some kind of laws. Using AI to fool people, to create music, impersonating and other things, can cause some major setbacks in the long run. I mean granted, I did use chat GPT twice and that Snapchat AI thing but that is just… no too weird for me.

It's not nearly as existential or terrible in its current iteration as its heaviest detractors would have you think.

It's also not as capable or all-encompassing as its heaviest supporters would have you think.

The way the stuff works currently is not "simulated brain thinks of answers" but rather "when you give it a prompt, it formulates potentially hundreds of google searches into a coherent sentence structure."

This sort of data processing is very different from what we've seen before. It does have a lot of uses, and will definitely have an impact on data entry jobs. Companies are going to end up finding out the hard way that the tech to have this stuff work efficiently and properly just won't be there for decades to pull off this layoff thing they seem to think is suddenly possible because the market won't functionally allow for it.

Artificial Intelligence is one thing. That's when something is programmed simply to react to a human action in a way that mimics behavior. When you message GPT, the detail and quality of the answer you receive is based upon the information that you give it in your prompt, and when you do that, you can generally see a lot of the same phrasings that you used in your question spit back into the answer. In addition, you may also think of things that are missing from the answer, or may even have contradictory information to what you're trying to gather details on. There is no "authority" on knowledge or information, and AI is neither suited for delivering consensus or determining what is the right thing to convey.

What people often think of when they hear AI is actually what people know as AGI, or Artificial General Intelligence, which is the point at which the centralization of information begins to resemble sentience. Once we reach the point where AGI is even blossoming, I'd get ready to hide from T-1000 units. We won't be there for quite some time, though.

What will happen is an increase of precision and severity of drones and military weaponry. Gone are the days of Obama "accidentally" bombing a wedding every week, here is the age of robot dogs with mounted assault rifles being guided by military engineers, aerial drone strikes that can be delivered en masse worse than carpet bombing, and logic processing that allows for faster movement. Augmented reality tools like Microsoft's Hololens are already being used by the military to create a head-up-display for soldiers and will likely be equipped with some form of GPT.

That's also if OpenAI doesn't go bankrupt over the next few months. The operating expenses for ChatGPT are at $700,000 a day, and it's expected to run them out of business by next year.

A good quote I've heard, but not sure where from: "we've created a technology that's only really good at convincing you that it doesn't suck"

I don't think it's crappy tech, I just believe it's unreliable.

Get what you can from it, it has a lot to offer, but like anything, no matter who makes it, it will never be impartial, it will never be a totality of objective fact, and it will never be intelligent. We need to make sure people, especially the elderly, are able to understand the "tells" to avoid getting scammed, sorta like Blade Runner.

I'm actually using ChatGPT for work, and it kinda sucks. My supervisor tasked me with creating SAT-style questions for the new digital test (since there are only 4 official practice tests available), and he recommended using ChatGPT to create the questions. It works ok for certain questions types, but it is absolutely abysmal for others. I've also noticed that it recycles the same few topics over and over again. The writing style is also exactly the same every time, which makes it hard to create Writing & Language questions, since it never varies its sentence structure or punctuation.

You also have to be extremely specific about what you want to get an output that's even remotely usable. "Create an SAT-style question" isn't enough, because it has no idea what that means. And of course, even if you are ultra-specific abo0ut what you ask for, the output will still need to be modified by a human to make it more representative of what the SAT actually tests. I have produced over 80 questions so far, and not a single one has been good to go straight from ChatGPT. All of them have required modification, most of the time to a significant degree.

tgags123: I've tried using ChatGPT for a few idea generation stuff and it's been terrible but I started using the new Twitter one for a while to see what it's like and it's much better.

If you're just doing thing for text generation, ideas, or it's something you have to manually change afterwards, I think it's okay but the stuff with AI art and voices have seemed to get even scarier since.

zanderlex : It's so odd you've found it awful. I've used it from creating lesson plans and practice scenarios for clients, to helping be write a friend's resume. Oddly, the one thing I find it's awful at is math. I was trying to use it for a business class I was taking, and it would get interest calculations wrong all the time. As a computer, I would assume the opposite.


tgags123 : That's so odd! I use it for multiple choice question creation all the time. I find I have to have the prompt well worded, and I always ask it to create far more than I need, so I can pick and choose the best ones. I'll usually write something like "Create 20 multiple choice questions, each with four possible answers, for College-level anatomy and physiology students studying the topic of the respiratory system and related pathologies" and out the 20, I'll probably find 5 I like. But that's 5 I didn't have to write! Also, it's good at creating them from a text. So I'll take a large section of a textbook, and paste it in the chat box and say "Create 20 questions from the above text with these paramotors blah blah", and it I'll probably use over half of those.