Current Events > What do you think AI is going to look like in the next 5 years?

(+) Yeah! (+)

I am quite cynical personally but I realise not everyone is.

Honestly, I don't know.

Deepfakes everywhere. What even is reality anymore?

It's already terrible. It'll only get easier to use and more believable.

kind9 posted...

Deepfakes everywhere. What even is reality anymore?

On top of a lot things being broken beyond repair because those in charge put all their resources into AI so much that even if the thing they put the AI in are worse off now they will refuse to go back to human labor because of sunk cost fallacy and because the people suffering for it won't be them but the poor powerless masses anyway, which plays into their innate sadism towards those who aren't wealthy

What could be a useful tool has already been twisted and corrupted into something that could seriously kill people.

Just scrolling down Facebook you'll find deepfake videos passed on as real content to generate clicks. That's the tip of the iceberg. What's stopping propaganda ministers from using similar methods to drum up public support for a war or genocide? Or even on a smaller scale, as AI porn gets better, you could create a video of someone innocent in the community raping another person or even a child. Has there been a public lynching yet as the result of AI? I know some countries have lynchings just on rumors!

Speaking of children, have you guys been to some of the AI chatbots sites? Just full of pedos creating fantasies without any oversight whatsoever in many circumstances.

It will help humanity

I don't think it's going to completely change our lives, but it's certainly going to make them more shit

AI is going to only be used by the wealthy who can afford the expensive subscription prices once the investment dollars stop rolling in and they need to start actually making money.

Best case scenario is that AI is a bubble and it pops.
Worst case scenario is that we all die.

uwnim posted...

Best case scenario is that AI is a bubble and it pops.
Worst case scenario is that we all die.

Financially possible with these big companies sinking everything in so fast, but the tech is here and it's here to stay.

I started using stable diffusion locally in it's early days and played a lot for a LONG time, and said back then that we were going to see exponentially crazy bursts of advancements

None of the above.

It will look more or less as it does now, with some refinements.

Error1355 posted...

AI is going to only be used by the wealthy who can afford the expensive subscription prices once the investment dollars stop rolling in and they need to start actually making money.

You can literally get free open source AI right now. Why would that change in the future?

kind9 posted...

Deepfakes everywhere. What even is reality anymore?

This mostly, at least as long as humans still have control of it.

Considering digital technology, especially a (self-developing) AI develops at an exponential rate, the moment AI passes the same capability as a human brain will pass by like a train racing past a station.

After that, I honestly can't even guess what will happen.

https://gamefaqs.gamespot.com/a/forum/5/5046cc23.jpg

Living in the country with no internet gets more appealing everyday.

ChatGPT can already control a web browser. It's called "agent mode".
At the very least, I believe you will be able to run an agent on your computer and have AI control your entire computer instead of just a sandboxed web browser.

Other than that, I dunno. It will probably be doing what it does now but a little better, along with whatever Agent Mode will allow it to do. Neural processors built into CPUs will allow AI to run on more devices locally without an internet connection needed.

The bubble will burst, but it won't make AI vanish off the face of the planet.

Not really any of those. Its uses will be expanded and some cool news tools will be created. The bubble might burst in the next couple years and it'll just become a standard piece of technology no one bats an eye at. Maybe not in the next five years, but fairly soon some legislation will start being pushed to mitigate it being used for disinformation and the impact of the server facilities.

Can I choose B and C?

I think the bubble will have popped and reset to a normal level. I think the efficacy of these AI models will be basically about what they are now, a sort of helpful tool in some cases, which still won't be 100% accurate.

In the west likely nothing, as its mostly stagnated and were just waiting for the bubble to pop. Whenever the government cant stop kicking it down the curve or something else occurs that forces it.

China though is another story.

Enclave posted...

Can I choose B and C?

I feel like it can't really be both of those at once, if it's a dead end then it's clearly not powerful enough a tool to also make massive changes like that

Governments will use it to create propaganda against opponents and to control narratives.

I give it one major reaction to a deepfake to make things go bad.

kirbymuncher posted...

I feel like it can't really be both of those at once, if it's a dead end then it's clearly not powerful enough a tool to also make massive changes like that

I think you underestimate how many companies are already saying "Eh, good enough" when it comes to say replacing the vast majority of their customer support.

It doesn't need to be good to change the structure of our lives for the worse. Hell, look at vibe coding, it's not going anywhere, there's going to continue to be people making apps via vibe coding and they'll have terrible security, bam yet more shitty AI making our lives worse.

We've only seen the tip of the iceberg when it comes to the bullshit people are going to do with current shitty LLMs.

ai123 posted...

None of the above.

It will look more or less as it does now, with some refinements.

That is bubble pops.
The major AI companies are trying to create ASI and thats why so much money is being poured into it. Either they succeed(and we are probably fucked) or the bubble pops.

I'm hopeful that we come up with actually decent use cases for it.

This is, of course, in addition to it continuing to be used for evil. That's a given. But I think it'll actually have the reverse effect of the common fear. As long as it's "innocent until proven guilty," I find it significantly more likely in the long run that more cases are dismissed on grounds that the evidence "could be AI" than actual false convictions using NN-generated evidence.

But it's also possible this doesn't happen at all. There were times when Photoshop was crazy good compared to people's ability to discriminate it, and we more or less did fine on this front.

And on the third hand, nothing is going to save us from the robots taking over all the jobs.

Vegy posted...

It will help humanity

At the cost of something

Hopefully it's positive

It's going to pop in the next 5 years. There simply is not enough energy production for these ludicrous power demands.