A lot more work needed in AI

[ReidG]

https://www.theregister.com/2023/03/02/chatgpt_considered_harmful/?td=rt-3a




The author says:
It not only told everyone I died but tried to fake my obit. Are we ready for this machine-driven future?

We will find good uses for the technology but it is still very early days and there are a lot of serious issues to be resolved with the stochastic parrot.

[AdamJL]

that's a given. No new(ish) is perfect on day 1.

[ReidG]

https://nymag.com/intelligencer/article/ai-artificial-intelligence-chatbots-emily-m-bender.html


Sometimes what has been promised never actually happens too. I still want my atomic car promised when i was a kid.:lol:

[AdamJL]

^ that's one of the... weirdest articles I've read, and I forced myself to read it all.



they will hardly unravel society....



The whole article read like that quote.

[Stonius]

When Fats Domino died in 2017, I was like 'I thought he was already dead?'


Point being, we can't judge AI poorly for making mistakes, because that is one of the defining qualities of humans. We all make mistakes.


Sure it's making more than you'd expect, but look at how far AI has come in the last few years, and extend that trend-line into the future. Those little niggles will get ironed out, and quickly. No it won't end society, in the same way as the internet didn't end society. But it will be a huge disruption and restructure the way we live. it's already doing that because it's already good enough at some tasks to make some jobs obsolete.


Like any new technology it will bring both enormous benefits and enormous problems, just like the internet did.


Personally, I think a properly trained AI will eventually do a lot better than most GP's in assessing patients and taking appropriate action.