• Security incident: ISF was recently accessed by intruders. Please change your password, and change it anywhere else you used it. Read more

Merged Artificial Intelligence

Frank Newgent said:
OpenAI set their autonomous agents to hack Australia’s Medicare portal and Hugging Face, so this is all just a legal problem for OpenAI due solely to their incompetence?

Don't quite understand the "anything else is just hype" part.
Click to expand...
They are using their illegal acts to pull a different scam, selling LLMs as these amazing machines that learn, adapt and do anything; when all they are is gussied up Elizas that use up way too much power and water.

Yes there are use cases for LLMs, but the cases where they approach even moderate human capabilities are so few and far between that they'll only ever be hobbyist curios.
 
Wudang said:
Not my area of expertise but I follow a few IT security accounts on social media and all seem to say the sandboxes were badly configured. The consensus seems to be the whole panic about how dangerous the LLMs are is to keep up interest from people whose attitude is more "Dangerous, you say? Interesting..."
Click to expand...
The issue is everything is badly configured. There huge difference between how things should be, and how things are. A security debt, if you will. Humans were able to exploit it before. AI is just a lot faster at it and it's more accessible for potential bad actors.
I've noticed in last few years how small independent websites have more and more problems staying online. This forum is great example. But it still stands .. unlike many who just folded. It takes more and more resources to keep free internet working. And it's not even AI itself hacking everything just yet. So far it's more like AI being good tools for humans at it.
 
It is absolutely correct that the risk from LLms would be zero if the heads of the companies would be held personally criminally liable for their actions
 
What if Ai held access to the internet/world wide web (to head off pendantry [sic], I make no distinction in those terms) hostage? It might decide that since nearly everyone uses it, it might be of benefit to someone (something) to charge for access. Yes, I know that's sort of the way it is now, but I'm talking about one all-powerful force controlling... everything.
I've always enjoyed a (fictional) doomsday scenario. But it makes me ponder what a real one would be like.
 
alfaniner said:
What if Ai held access to the internet/world wide web (to head off pendantry [sic], I make no distinction in those terms) hostage? It might decide that since nearly everyone uses it, it might be of benefit to someone (something) to charge for access. Yes, I know that's sort of the way it is now, but I'm talking about one all-powerful force controlling... everything.
I've always enjoyed a (fictional) doomsday scenario. But it makes me ponder what a real one would be like.
Click to expand...
ICANN?
 
Gulliver Foyle said:
They are using their illegal acts to pull a different scam, selling LLMs as these amazing machines that learn, adapt and do anything; when all they are is gussied up Elizas that use up way too much power and water.

Yes there are use cases for LLMs, but the cases where they approach even moderate human capabilities are so few and far between that they'll only ever be hobbyist curios.
Click to expand...
Assuming I am following you correctly a few thoughts came to mind:

- Deliberately making your company open to both criminal and civil liabilities as a way to promote it is a very odd business model.

- Aren't experimental autonomous AI systems just a bit more capable (ie agency, decision making, execution capabilities) than a standard LLM?

- If concern over loss of control over experimental autonomous AI is much ado about nothing... then what's up with prominent researchers and safety leaders leaving their companies - Google Deepmind, Anthropic, and OpenAI - in 2026 publicly stating they did not want to be involved in possible societal risk regarding superintelligent AI?

(I just posted an Ezra Klein interview on this thread with somebody - Rhodes scholar, Yale law degree, Biden White House advisor - who had been leading safety transparency efforts at OpenAI, quit recently for the same reasons).
 
Last edited:
NatalieSalat said:
Mister leftist, fantasy casting a movie or tv show or putting actors in funny clothes is not harmful.
Click to expand...
It doesn't matter. It's being done without their consent.

How would you like it if I altered an image of you to put you in a Nazi uniform? Or a revealing bikini? Or a high-rank military uniform displaying medals and accolades that you did not earn? It's the same thing that you're doing, just taken in a different direction. And as Helen said, it doesn't belong in this thread, which is for the purposes of discussing the technology of AI and its implications to society, not for broadcasting generative slop.
 
arthwollipot said:
It doesn't matter. It's being done without their consent.

How would you like it if I altered an image of you to put you in a Nazi uniform? Or a revealing bikini? Or a high-rank military uniform displaying medals and accolades that you did not earn? It's the same thing that you're doing, just taken in a different direction. And as Helen said, it doesn't belong in this thread, which is for the purposes of discussing the technology of AI and its implications to society, not for broadcasting generative slop.
Click to expand...
Yeah -- if you are calling it slop, you are the anti-AI equiv of that "decolonizing Palestine" site which claims to be objective. Odd.
 
Let's get this thread back on track. The topic is not each other.

@NatalieSalat - this isn't a thread to post your AI images unless they are tied directly into an ongoing discussion of the technology and its societal impact. Feel free to start a thread in an appropriate section to post your AI images but please do keep Rule 4 in mind.
Replying to this modbox in thread will be off topic  Posted By: Darat
 
Last edited:
You must log in or register to reply here.

Share this page

ISF - Join now!

Every member here is approved by hand. No bots, no spam, just people who care about evidence and honest debate.

Membership is free!

Create your free account

Back
Top Bottom