Dusty, Good information. I did similar search on Perplexity (not pro version) and got different answers depending on how question was phrased. Sources were discussion forums similar to ours!
Ultimately the accuracy of any information goes back to the source. I still remember years ago doing some medical research I found a claim in an article or textbook and then looked up the original source. That source did not support the claim.
I wonder if there’s any consensus about usage of AI on the forum? We (including me) do lots of conjecture on the forum because a lot of info just isn’t available. I’m fine with anybody expressing an opinion of what they think.
However I’m less thrilled with people either using an AI generated claim as coming from themself or citing it as a source without any attempt to verify the claim, or even worse making claims about how reliable AI is.
I put this in Uncategorized category as didn’t see an appropriate category. I was hoping for something like “Forum Decorum :>)” Basically what do others feel is appropriate use of AI on the forum. Can we agree on any guidelines?
Can we agree on anything? Ok, I agree with you. It’s definitely easier to ban it or discourage it now than to try when it’s accepted as part of the fabric of the forum. I see forums in general as havens of human discourse.
Exactly. Worst example I’ve seen was a definitive conclusion where the reference was merely an editorial. So for my purposes so far, AI is just a tool to speed the collection and parsing of search results. I think it’s great so long as the user understands the limitations.
Good luck with that. The single most egregious thing I can remember reading on this forum had nothing to do with AI. If I could write a guideline to prevent that one it would be “no posting within 8 hours of your last drink.”
@Abram_D , @rasmus_braun Does forum management have any opinion on use of AI on the forum? It seems like people are increasingly using it and some have incredible faith in it’s accuracy.
I don’t have any problem with posting an AI response as long as it is clearly identified as an AI response. We have had at least one member who has posted useful information gained from an AI inquiry but he always labels it as such and names the AI machine in the first line.
Very good response. The AI tools have their limitations just like any other software (ever had to reboot your computer due to hung up software?) so we should be cautious about relying on these tools and identify when the info was obtained through AI.
Then, OTH, there’s this article by a NY Times Tech columnist who studies AI for a living (I’ve shared the article, so you should be able to access it without a NY Times subscription):
Three arguments for taking progress toward artificial general intelligence, or A.G.I., more seriously — whether you’re an optimist or a pessimist.
Neither the TechSpot author, the Tow Center for Digital Journalism, nor the Columbia Journalism Review sound unbiased, e.g., the CJR footnote at the bottom of the Tow Center article:
Has America ever needed a media defender more than now?
I think they’re all going to find the train is leaving the station, and they’re basically arguing that Henry Ford’s Model T is not to be trusted as a mode of transportation. Why, most of America isn’t even suitably paved for automobiles! Best to stick with blacksmiths and horseshoes…
I haven’t tried Perplexity AI. I’m a ChatGPT Plus subscriber, and as I’ve noted on the forum, I always crosscheck the answers for anything that matters in a number of ways, including trying other AI models, using Google/Bing Search and reading hit links, consulting human experts, etc.
I’ve been very pleased with ChatGPT results, especially in areas where I have some expert knowledge. I basically agree with the NY Times Tech columnist on where we’re heading (and where we’re at now).
A known difficulty with AI is actually called “hallucinations”, by some folks who are familiar with the technology.
My personal preference is to not allow AI “findings” to be posted on this forum at all, even if identified as such, because of this accuracy limitation. YMMV, of course.
Thanks for all of the sharing of actual life experiences by so many humans!
Best, Bob.
