Online Forums in a world of ChatGPT?

[wtr]

All,

I'm curious, what happens to a forum like this with chatgpt and other AI tools that can essentially trawl through the web and report results to questions?

Is there ways we can be symbiotic with chat AI's?

Could they be used to prevent duplication of questions?
Could they be used to lessen the burden of administration of messages/rules infringement?
Can we get the suppliers of AI to dry run this on EDA?

Regards,
Wes

 

[FvM]

How can you expect ChatGPT to collect reliable solutions from the internet? We see many examples that neither humans are able to.

 

[KlausST]

Hi,

Is there ways we can be symbiotic with chat AI's?

I guess it already is. I guess some of the ones who ask questions here use AI, as well as some of those who answer questions.
And I´m quite sure in future we will have AI as a direct "forum member".

Could they be used to prevent duplication of questions?

I can´t see how. People ask questions. Some of them ask questions that have already been answered many times in this forum.

***
AI in forums generally:
A forum is not meant to help people in first place. A forum is a profit centre.
The operators of a forum are interested to earn money. So if AI increases the income, the operators will use AI.
This forum is supported by unpaid moderators. So financially the moderators are non existent.
I guess using an AI in a forum also means the forum operators need to pay for the AI.
So it´s an economical question whether and how AI is used in a forum like this.

Klaus

 

[betwixt]

AI 'learns' by reading other peoples (or AI generated content), it can only be as accurate as the content it sees or derives from it.

The moderators here are all experienced in electronics and possibly also in manufacturing, we learn from reliable, trusted sources and from our own investigations. It gives us breadth of knowledge based on real situations, and (speaking for myself) the reasons things sometimes fail. It gives us intuition about what is correct and sensible and also how to politely guide those less experienced towards a satisfactory resolution to their questions.

AI, at least at the moment, lacks the ability to differentiate between what it scans on the internet and what an engineer would dismiss as incorrect. If a thousand internet sites said "The moon is made of cheese" the AI would make it a trusted assumption and tell it to others. This message has probably already been scanned so make that 1001 sites.

I shudder to think whether AI scans some of the YouTube sites, there are so many fake electronics sites they couldn't be counted. There are some very good ones too but AI wouldn't see the difference, only the content. So many video sites show total nonsense that AI would add to its repertoire, they are all 'click bait' to bring advertising revenue to the poster but only someone with experience would recognize that as the case. See how many free energy videos you can find that glue spark plugs to the back of loudspeakers to create perpetual energy and you will see what I mean.

Brian.

 

[KlausST]

If a thousand internet sites said "The moon is made of cheese" the AI would make it a trusted assumption and tell it to others.

I guess this will be one of the future tasks of the AI algorithm. The problem is that when a thousand sites say so .. then AI will repeat it, how often? And every time it repeats it it will become even more trustworthy for an AI. I´m not sure how this can be stopped.

It is similar to how propaganda and racism work. In my opinion, only healthy skepticism and education helps.


Klaus