How Dangerous is AI?

[pablolie]

I wonder if this is likely to happen anytime soon. Let's rule out omniscience or whatever, but just say that the AI has to do as well as a smart, unbiased human in establishing whether a bit of information is true or not.
...

That is impossible because the models are entirely trained by humans for a very focused application case.

ChatGPT ... for many corner cases, we train it collectively. Because the model does not have a lot of data yet. You can argue with ChatGPT if you go for an obscure question, and debate with it. Ask it something like "When was the earliest Ming Dynasty jade vase made?" and argue with it suggesting any bizarre data point, and it'll start using your input until someone else corrects it. AI has zero intelligence on its own. It's basically specialized automation with a big data lake.

 

[kemmler3D]

That is impossible because the models are entirely trained by humans for a very focused application case

Acknowledged that current GPT AI models don't work like that, but "AI" is a very large category, and there's no real reason software that follows different heuristics to evaluate the quality of information couldn't be built independently or added on top of GPT-type software.

 

[pablolie]

Acknowledged that current GPT AI models don't work like that, but "AI" is a very large category, and there's no real reason software that follows different heuristics to evaluate the quality of information couldn't be built independently or added on top of GPT-type software.

ChatGPT is generative AI indeed. They did pick their market niche in text generation. Other frameworks go for other application cases. Yes, using the same DNN principles and convolutional or recursive models, but ultimately you need to specialize to have a successful business case. OpenAI clearly doesn't think composing music is how they'll make money. They invest in generating text and their focus is on being the best in that narrow niche. They are busy enough with Google going after their business case.

AI is no longer a science experiment, it is a billion dollar biz and the investment guides it.

 

[Anton D]

It's so dangerous, it buys records and listens to music using a turntable.

 

[pablolie]

It's so dangerous, it buys records and listens to music using a turntable.

It is so dangerous, it is re-mastering Spyro Gyra albums on its own and posting them on HDtracks for a $20 premium. :-D

 

[pseudoid]

Uze awl do know that everything uze guyz say everywHERE is being 'logged in'.

If someone in the future asked it:

It's so dangerous, it Do you buys records and listens to music?

It may have to answer "Yes, I do...

...using a turntable

And you @pablolie,... ummm... SpyraGyra was a good pick but... your statement becomes...

...so dangerous...

 

[fpitas]

Finally, an AI article that makes sense:

MSN

 

[fpitas]

You may resume your normal panic.

 

[tw 2022]

Finally, an AI article that makes sense:

MSN

interesting... if it can be corralled , it seems like a useful tool in many areas.. if ai becomes self sustaining all bets are off.. the problem is *WE DON'T KNOW* as a society where the tech stands from day to day and we won't until the shit hits the fan (or doesn't).. nice article if this info is accurate and not a political stunt to keep ai unregulated .... my thought is tech *will kill us* ... social media is the devil on earth, 100 times worse than any drug problem in the history of ever...if it was up to me all tech targeted to public use would be highly regulated...

 

[fpitas]

interesting... if it can be corralled , it seems like a useful tool in many areas.. if ai becomes self sustaining all bets are off.. the problem is *WE DON'T KNOW* as a society where the tech stands from day to day and we won't until the shit hits the fan (or doesn't).. nice article if this info is accurate and not a political stunt to keep ai unregulated .... my thought is tech *will kill us* ... social media is the devil on earth, 100 times worse than any drug problem in the history of ever...if it was up to me all tech targeted to public use would be highly regulated...

Not by coincidence, the same billionaires who champion AI are responsible for most of the social media debacle. It's time people realized these guys do not have our best interests in mind. For some odd reason they have all become folk heroes.

 

[levimax]

The issue with AI, just like with the internet before it, is that it's main use is to spy on people. AI just takes it to the next level and automates it. While many argue this is great for the convenience of the users and not to be concerned about privacy I always ask them "Name one example in the history of the world where being spied on worked out well for the person being spied on" and so far no one has come up with an example.

 

[NTK]

The issue with AI, just like with the internet before it, is that it's main use is to spy on people. AI just takes it to the next level and automates it. While many argue this is great for the convenience of the users and not to be concerned about privacy I always ask them "Name one example in the history of the world where being spied on worked out well for the person being spied on" and so far no one has come up with an example.

The internet? You'd rather do without it now?

 

[tw 2022]

Not by coincidence, the same billionaires who champion AI are responsible for most of the social media debacle. It's time people realized these guys do not have our best interests in mind. For some odd reason they have all become folk heroes.

exactly ,, to put it in a nutshell , i believe these asshats are trying to rule the world through misinformation... what i find *extra* appalling is that much of the funding for the tech is taxpayer funded through grants and grift courtesy of congress ... all the while vets, seniors and disabled folks gotta fight with the addicted and urban campers who suddenly "see the light" for affordable housing(i have no quarrel with either group , affordable housing should be a right in this country based on communal wealth) .. i'm(disabled) on so many waiting lists for housing i literally can't keep track of it all.. and everything is *years* on a list... all the while Elon Musk is funding his next hostile takeover with tax subsidies... it's criminal...

 

[tw 2022]

The internet? You'd rather do without it now?

uh, you don't think regulation of social media might help? not to mention data mining.. this *either /or* " argument is rather , shall i say "grade school level " argumentation...

 

[levimax]

The internet? You'd rather do without it now?

No I never said that and at this point I could not make a living without the internet. That doesn't mean that I am not being spied on 24/7 and at some point that may end up being sub optimal for me.

 

[NTK]

uh, you don't think regulation of social media might help? not to mention data mining.. this *either /or* " argument is rather , shall i say "grade school level " argumentation...

If you are so good at argumentation beyond the "grade school level", don't just say "you don't think regulation of social media might help? not to mention data mining".

Assuming your answer to your rhetorical questions is yes, be specific of what regulations you are proposing: What are you going to regulate? How are you going to regulate? And who are you going to regulate?

 

[tw 2022]

If you are so good at argumentation beyond the "grade school level", don't just say "you don't think regulation of social media might help? not to mention data mining".

Assuming your answer to your rhetorical questions is yes, be specific of what regulations you are proposing: What are you going to regulate? How are you going to regulate? And who are you going to regulate?

good answer: actually you brought some "game" with this question... what regulation am i proposing ? limiting agreement contracts to curb data mining... these web sites don't have the inherent " right" to assets based on "just because"... it's the wild west on the internet.. regulation resembling over the air broadcasts can be imposed pertaining to social media , where licensing is based on accountability(truth)... i'm already anticipating the "govt over reach" argument here .. regulation hasn't killed tv , the deregulation of cable outlets has... shall we continue?

 

[kevinh]

Big Tech Companies working with Governments to censor information is the death of freedom. The Ai will present us with a Narrative rather than information.

 

[slaweks]

i can see that happening , i heard a recent interview (can't remember what source, but most certainly "mainstream" based on my info sources) that predicted that the 2024 election will be the first test on the effect of AI as it pertains to the ability foreign govts to interfere with the US election in mass quantity.. iow , the russians and koreans and chinese are planning on flooding social media with misinformation courtesy of AI... good times

Foreign actors influencing elections is nothing new. The best way to detect and deal with it on social media is AI.

 

[egellings]

The issue with AI, just like with the internet before it, is that it's main use is to spy on people. AI just takes it to the next level and automates it. While many argue this is great for the convenience of the users and not to be concerned about privacy I always ask them "Name one example in the history of the world where being spied on worked out well for the person being spied on" and so far no one has come up with an example.

Well, the purpose of spying on someone is to grab their money when they are not looking.