Site policy on AI generated posts?

some teachers refused to use electronic calculators and persist on the logarhytm book. Completely absurd and stupid thing.
Did all students at that time have the electronic calculators?
I have grown in 70-ies in a similar situation - but electronic calculators in USSR were very rare then so the rule of not using them was providing all pupils/students the same basis.
 
Yes, we have the popular Casio FX120. But I have a smaller one, one of the firsts with LCD display. I still have it.
 

Attachments

  • 17191736617481819467471496983769.jpg
    17191736617481819467471496983769.jpg
    297.4 KB · Views: 13
Member
Joined 2021
Paid Member
I think it’s important to contemplate a forward-looking policy on this.

There is a spectrum from terrible post to excellent post.
No one needs AI to make a terrible post (humans can make **** posts easily on their own) and excellent posts remain excellent no matter how they came to be (using one’s vast experience, quoting work of the masters, tinkering with LTSpice or soon, most likely, working through a problem assisted by AI).

So, most likely, it should be enough to ask people to declare “post made with AI and unverified”.

Of course, using AI to create spam or scam people should be punished using already existing rules.
 
  • Like
Reactions: 1 users
Founder
Joined 2000
Paid Member
We also don’t want AI generated content polluting search results.

A new BBcode tag similar to the code or spoiler one would work, but still allow post corpus pollution.

Requiring AIGC to be attached and referenced as a clearly labeled text file might do the job for now, short of a total ban. Posts can refer to what’s in the text file but no AIGC in any post itself. Linking to off site pastebins invites link rot which is also unhelpful.
 
Founder
Joined 2000
Paid Member
Indeed.

There will always be bad actors. There are lots now, and there will be even more in the future. All we can do is our best. And that is to come up with a pragmatic defence plan and iteratively improve as time goes along.

I mean the following in no negative way, because if it's true we're doing our job well. I imagine most members would be blissfully ignorant about the multiple levels of protection the site has to have in place currently to defend against the swarms of daily infiltration attempts to the site powered by the current level of technological maturity of bad actors.

Aside from technological protections on the way in to becoming a new member, the mod team manually and carefully inspect every single application for registration, as well as new members' posts. The result is that thanks to their very hard work 24x7, combined with some good technology to assist, diyAudio is the great place it is today, free of spammers and scammers. We're lucky our site is so niche and technical - it makes it very hard for spammers to present themselves as genuine DIYers.

"Today's AI is the worst you'll ever use" - David Friedman. I think those laughing off AI are somewhat living in denial, and focusing on what it doesn't do well today, rather than what it does do well. It will get better and better at the things it doesn't do well.

One day we might have to require "Human ID" to register at diyAudio, or something else that's come into existence almost certainly by that time, to solve the common problem of sites (and governments) needing to verify a human is registering. One day we might have to charge $1 to become a member (no spammer wants to pay, they just want to pick the lowest hanging fruit). Or require real human vouching, like couchsurfing.com used to do.

We'll deal with those challenges and adapt our method, policy and technology as time goes on.

Until then, I think diyAudio should be a place for us humans to interact and enjoy our hobby. No diyAudio members should have to wonder if they are talking to another human being.
 
  • Like
Reactions: 3 users
What exactly makes you post a statement like this?
Could you elaborate what causes you aggravation?
Like which one?
I think it seems clear, but if you really want to be so "aggressive" I'll reply that it appears (very) evident that self-declaring that something we are treating as one of the most difficult issue to identify (and therefore to be resolved) is "simple for me" makes no sense, from my modest point of view.

Firstly because everyone is good at making positive self-declarations about themselves, secondly because we were talking about something completely different and the solution to what others do that is unbecoming is not to self-indicate what is positive you are doing and what you think others should reproduce on turn.

However, neither your comment before nor your comment now bothered me in any way, I replied what your comment aroused in me and I don't understand what is causing you so much "aggravation" (whatever you call it).
This seems to me very close to trolling.
 
it appears (very) evident that self-declaring that something we are treating as one of the most difficult issue to identify (and therefore to be resolved) is "simple for me" makes no sense, from my modest point of view.
Ok, I reserve the right for a different opinion.
I do not see any reason to continue our dispute so you have your way and I have mine.
 
  • Thank You
Reactions: 1 user
People have limitations, such as, a finite patience. Bots can be made to emulate an endless patience. Learners, beginners and experts alike can only gain from AI powered bots.

I dare suggest, to replace the search function of the forums with an AI powered bot. It will simply make miracles. However, AI requires powerful hardware, such as graphic cards and tensor processors, especially, when it is being trained.

AI can also be made to moderate posts to filter posts containing abusive content. This is a powerful addition which will make the forums more friendly to non-abusers.

Give it time, AI is bound to improve. There is too much to lose by opting out of AI powered searches and moderation.
 
Last edited:
  • Like
Reactions: 1 user