New Movie on AI downsides

[YB88]

A movie was released at the end of March on the topic of what AI is doing to --- everything (social, economic, political etc etc) --- and thinking through the downsides as well as upsides. This is:
The AI Doc: Or How I Became an Apocaloptimist

I've not seen this yet, though here's a preview:
https://www.youtube.com/watch?v=yvq-jYira9I

What I did do is listen to an equally long podcast by Sam Harris discussing these topics in, I think, more depth and with more intellectual and less emotional content, with Tristan Harris. I don't know that I'll bother with the movie, but found the podcast discussion worth the time. This can be seen (for free) on Youtube:
Escaping an Anti-Human Future: A Conversation with Tristan Harris

I use AI a lot and have been trying (in my amateur way) to understand the underpinnings of what makes it work and trying to keep up with all the many rapid advances.
So I guess I have somewhat of a built-in "pro" bias towards AI, but I think it's important to think through the downsides and likely outcomes as it reshapes our world.

Tristan Harris does an excellent job of presenting the dangers, the downsides, and how they relate in a non-equivalent way to the various upsides.
Until listening to this my thought on this whole topic had been "If we don't do it, the Chinese will, there's no stopping it so we should just do our best to mitigate the downsides as we go along". I'm a little less sure of that now. I don't necessarily accept all of Tristan's premises, but this will have a really big impact on humanity's future, so it's worth thinking through this stuff. Even where it's uncomfortable. Even when some of the hyperbole seems too ... hyperbolic.

Bottom line is that I recommend the podcast, not sure about the movie.

 

[COcheesehead]

Most of ai still requires a critical eye. If you are not careful, it will tell you what you want to hear, On a positive side it can expose you to ideas you may not have thought of and create a path to better understanding or communication.

Some we know have harnessed it to make their jobs easier. Many of my wife’s fellow professors use it to create test questions, write presentations and cut busy work out of their days. My wife just used it to negotiate buying a car.

 

[sengsational]

The problem with the world in general is that it's much easier to break stuff than it is to make stuff. So as AI becomes more ubiquitous and making RBT's goes into the hands of lunatics, the rest of society will pay the price. Oh, my acronym Really Bad Things.

One RBT in the hands of lunatics that we've already seen was airliner full of fuel. But that's small potatoes compared to a virus. There are tons of pandemic level viruses that have been published. Synthetic biology AI means anyone with the will could build that virus.

Another way to break stuff is to leverage all of the vulnerable computer systems. There are billions of systems connected and most of them will never be patched. AI can find vulnerabilities and write exploits hundreds of times faster than humans have been able to do it. So seriously breaking financial systems seems inevitable.

There's a thousand wonderful things that AI can do, but the lunatics are likely going to ruin it for us.

Have a nice day

 

[kgtest]

The problem with the world in general is that it's much easier to break stuff than it is to make stuff. So as AI becomes more ubiquitous and making RBT's goes into the hands of lunatics, the rest of society will pay the price. Oh, my acronym Really Bad Things.

One RBT in the hands of lunatics that we've already seen was airliner full of fuel. But that's small potatoes compared to a virus. There are tons of pandemic level viruses that have been published. Synthetic biology AI means anyone with the will could build that virus.

Another way to break stuff is to leverage all of the vulnerable computer systems. There are billions of systems connected and most of them will never be patched. AI can find vulnerabilities and write exploits hundreds of times faster than humans have been able to do it. So seriously breaking financial systems seems inevitable.

There's a thousand wonderful things that AI can do, but the lunatics are likely going to ruin it for us.

Have a nice day

IF AI can start checking in logical code and functions without a human knowing, nor understanding it...we could be in for some problems. I think that scope creep is already occurring...the human is the gate, but it's so so easy to look the other way, click auto-approve and let the exploits begin. Finding the issue is why war rooms were created in IT groups.

If we can't find the issue, say the filename and date look like they never changed so that code is never analyzed...we get duped...the problem becomes much harder and longer to solve for sure.