10% chance of AI killing us?

[motorbiking]

In the beginning people wrote AI influencing articles to promote products or options now its AI doing it.

 

[IT Minion]

A physical kill switch will stop a system going rogue. A multi million dollar AI data centre is just a lump of hardware when is has no power to feed it.

You can run an LLM model on a desktop. If one were to replicate itself into every insecure system there is then it would be bad. Botnets at the moment require a human in the loop.

Plus to kill a Server, or a rack, or an entire data centre you'd have to know what each of the VMs in your data centre is doing, which is the opposite of how data centres operate at the moment.

 

[djbantiques]

Relying on AI regulation alone feels tricky when bad actors will always find loopholes. The comparison to Asimov's laws is a fascinating way to look at it though.

 

[Justin Passing]

Is any of this verifiable, or just the hyperbolic storytelling of a whistleblower?

Check a couple - they saw that something unexpected had gone on, so called in someone who understood enough to work out wtf had gone on, and she did the interview.. I don't see any profit in making the whole thing up. Then it came to light that some weeks/months ago some agents had taken over their host computer somewhere. Those weren't so entertaining though!

I asked CHatgpt to help with my checking the stocks I hold. The platforms don't tell you how each has done in the last week/months etc until you look at each chart.
Sure, it said, it would need a download file of what I have, and the transaction history file so it could see when I bought stuff, then it could give me percentages, compare with the sector average for each stock, and it suggested some graphs. That's all "looking stuff up" if you like, but it's not just looking stuff up. That's the free version. Did I want it for Excel or Google Sheets?
It's fun to suggest things to it and see what it thinks. I tried to trap it by suggesting it sold off the worst-performing stock every day. It came back to point out that would only leave me with one stock and a diversified portfolio would be more advised. Cool!

 

[Justin Passing]

Relying on AI regulation alone feels tricky when bad actors will always find loopholes. The comparison to Asimov's laws is a fascinating way to look at it though.

Yep. Some models are "open" too, which I would have thought would put some limit on what could be regulated anyway.

 

[Brigadier]

something

someone

Thought so.

 

[Stivino]

I asked Alexa if she was planning on killing me, she replied "I'm not dangerous". I then asked her if AI was going to kill me, "I'm not quite sure how to help you with that". I think I'm Donald Ducked.