Hello community!
We recently have seen an uptick on members using ChaptGPT responses to answer members questions. There are often times when multiple “members” respond with almost identical answers and making it harder for our Champions to have visibility on their answers to these questions. How are you all managing the use of AI generated responses? Have you updated your T&C? Sent warnings or suspended members? Would love to hear how other communities are navigating this.
+7I faced this with one or two users a few months back and just talked to them directly about it; that stopped the behavior. But I know that I still have a lot of work to do something about it.
On my list:
But we also won’t outright ban the behavior because we’re developing our own chatbot trained on our documentation. (it seems hypocritical to outlaw AI use when we also provide it; plus someone who is skilled at prompting can be a helpful contributor).
My hypothesis is that when we integrate our own model to be the first reply (beta phase of our chatbot), it will reduce the behavior because it takes away the easy answer they can get via AI. The AI bot/agent will get all the low hanging fruit, leaving more nuanced and complex discussions for humans to chat about
Finally, when we get to that point, I will encourage users to be transparent in their use of the AI. Already a percentage of our questions are the “let me google that for you” equivalent on the community, sharing a previous post with the answer. I want the AI version of that to be “I asked the {company provided bot} and it said this ____. From my understanding, that should work, but you may want to consider ____”
Also this recent post may be helpful for analogous ways to monitor for AI content:
Hi
Using AI to assist with creating community replies
While using AI to assist with community questions is currently permitted, members are required to note it at the beginning of their reply. Something like "This reply has been created with the help of AI" As with any reply, members need to make sure it is applicable to the question.
+5This is a great and timely question. We had a similar conversation a little while back (though a little more in relation to the risk of spam):
+2This is a great and timely question. We had a similar conversation a little while back (though a little more in relation to the risk of spam):
This is something I started to think about when I saw this topic. Thanks for sharing
We recently have seen an uptick on members using ChaptGPT responses to answer members questions.
Are you sure they are not bots?
+2
Are you bot? No.
😀
Hiya!
Just a heads up, my answers come from more of a Super User/Forum Volunteer perspective as that’s the role I play. :)
For me, I tend to find AI generated posts to be more annoying than helpful and have a preference towards just getting the Forum Moderators to throw banhammers at the users who use such tools, unless the AI content actually adds value. I don’t recall seeing anything AI-specific in our T&C but we tend to treat it as spam which is covered, since almost all the AI stuff I’ve seen to date is basically spam or some other form of junk anyway.
However…
Given that Energy related stuff is annoyingly complicated and sometimes I have to dig pretty deep into the puzzle solving, I do sometimes use AI to help me with the research I do while writing posts or comments. If I need the help, normally I’ll just feed the question (or some other prompt) into Gemini, let it number crunch that and use whatever it comes back with as a starting point for further research but then do everything myself from there. I only copy/paste stuff directly back into the Forum very sparingly for something where the answer I wanted to write out is basically better written than I could do myself (and I’ve double checked it!) but I will always then tag it as having had AI assistance. This stuff only really makes up a tiny proportion of my content though as I find half the time, the AI messes up and goes off track anyway.
IIRC the only thing I’ve posted that has been majority AI written by Gemini so far was the instructions to do a specific gas meter test because I couldn’t find the existing guide on the Forum and I was feeling a bit too lazy to type the whole thing out myself for the 1,000th time so I just had Gemini do it (there’s only a small number of ways to write out the instructions for that test anyway!) and made sure the output matched what I know of that test. I also made sure to add a note to say that I’d used an AI tool to help write it out.
The biggest AI usage I tend to see though, is the most annoying stuff - spam. We see a fair amount of cases where some random user suddenly posts a weird comment that gravedigs an old thread with stuff that isn’t necessarily spam but definitely doesn’t help much. More and more, it triggers my AI detection tools whenever I do my checks...
For me, I’d say allowing your trusted users/super users/forum volunteers to use AI to help with their research for tricky questions is totally fine because that’s no worse than using Google to do research like we always did before all these fancy tools existed. I just prefer it when folks write out their own responses without getting a hivemind powered chatbot to do it for them. But I’d still suggest asking them to tag a note saying they used AI to help them with the research and set expectations to double check the answer first.
As for everyone else? That’s more tricky. I’m more likely to be OK with a particular piece of content if the member who posted it has clearly taken the time to double check everything, added their own thoughts and the post as a whole adds useful value to the Forum. If, however, it’s just an AI Spambot jamming up the Forum then you can be sure that SkyNet Blastoise here will whack it with the BanHammer of Doom 3000 before you can say AI. Ok, maybe not that fast, but you get the idea - I’ve got a good eye for that stuff!
If you ever had a profile with us, there's no need to create another one.
Don't worry if your email address has since changed, or you can't remember your login, just let us know at community@gainsight.com and we'll help you get started from where you left.
Else, please continue with the registration below.
Enter your E-mail address. We'll send you an e-mail with instructions to reset your password.
