Should AI generated posts be banned or otherwise regulated?

The more you use AI, specifically ChatGPT in my experience, the more you’ll find its answers to be long winded and not what you asked, sychophantic, blatantly wrong, and completely made up.

Example: I asked it to compare the Rogue RP-5 to the Rogue RP-1 deriving its answer from actual user experiences. Except I typo’ed and wrote RP-11.

It told me through real user experience, they found the RP-11 to be more refined and musical than the RP-5, with a blacker background and cleaner sound. For those who don’t know, there is no RP-11. This is just a quick audio example of its nonsense. I’ve myriad more unrelated examples.

Forgot contradictory!

@thecarpathian 

  that’s funny because I had typed the exact same query into Chat GPT a few months ago, minus the typo.  I just rechecked the answer and it is essentially the same as yours 

AI is fallible in its response to questions of those industries that are discretionary. Because AI lacks a sensory experience, it has to rely on the human experiences that it gathers in it's searches on the Internet.

There is no audio manifesto that curates responses for AI, and there are few absolutes in this industry outside of specifications (science), room dimensions (math), and what else? AI cannot compare components outside of what the manufacturer provides, unless it is doing so via sites where human feedback is curated and gathered. You can measure a component and make a distinction based on those scientific measures, but does it sound good? AI ain't got ears and has to rely on humans for such feedback.

Audiogon and similar sites are typically part of the results as a source for AI. These sites deliver real world human feedback on sound, something AI is not able to quantify, so it sources what it can on the subject, which is subjective. 

If you query Audiogon on a particular piece of equipment- and then query AI on the same piece- it is likely that AI will source Audioqon as part of it's findings. And if you are the one who wrote a review, you may end up being the source for AI's response even though you may have been winging it at the time, or just prematurely providing a review that as we all know...changes over time. It's why we give up gear to buy new gear, your discretion changes, as does your wallet, health, etc. Far too many variables for AI to understand. Maybe over time, but AI's response is still going to be validated by audio experiences of humans who write about what they hear.

IMHO, I think it likely that AI will be inundated with disinformation on the subject as free sites like Audiogon can make a difference in your promotion of a product and it may make sense for audio manufacturers to attempt at building a brand by infiltrating AI by taking on a larger role in these spaces where AI sources results.

@thecarpathian 

Congratulations, you taught AI that there was, in fact, no such thing as a Rogue RP-11. There was a similar situation some time ago with parkers thread where there was, in fact, for a certain period of time a Musical Fidelity A380. Oddly, like your experience, AI was, in fact, dyslexic. It has an affinity to mix up numbers, as there is, of course, an A308. 

AI will not tell you God exists, even though we all know that God does exist, right? 

AI says, "Because "God" is generally defined as a being outside of space and time, the question typically falls outside the realm of empirical scientific proof".