Skip to content
HearingTracker
Log in

Any chance we could introduce a forum rule about AI content

This topic is closed.

69 replies1.4k viewsLast reply

Jump
Link to this topic
Edited

Apart the fact that content burns more energy than a small star, it’s particularly unhelpful in terms of answering questions about the motivation, manufacture and fitting of hearing aids: not least because some of the most obvious answers are usually the inverse of what might be expected.

Worse than that; it leads to threads becoming unreadable due to the choking volume of bilge that it produces to ‘evidence’ the flawed arguments that it makes. It almost seems as if people follow it because of the volume of diatribe associated with the post rather than being able to properly understand the crux of the question.

BREVITY!

27 people in this conversation24 min readJump to latest

Three cheers! I absolutely agree. should not be a citable source. If one wants to use it for research, delve in and find the sources of information and cite them.

Fully agree.

No more -generated answers that are copied and pasted.

In reply to @MDB

Echo cheers back!

While being able to use natural language to ask a chat is impressive, it’s the incorrect information that it spews back that is problematic. So then you have to hone your question. Well then you may as well have done the googling anyway and going down rabbit holes yourself.

@Um_bongo said:

Worse than that; it leads to threads becoming unreadable due to the choking volume of bilge that it produces to ‘evidence’ the flawed arguments that it makes.

You do have a way with words that I enjoy.

I tend to use the word nonsense.

I completely agree! This forum thrives because there are real humans here. Let’s not ruin it with !

Totally agree. I actually use results extensively from Google. They can be very useful. But you have to know what you are getting, and how to use it.

And I agree, a forum like this is NOT the place for it.

I don’t know that I want to be that black and white about use. Perhaps the rule could be something like: “Don’t paste long AI outputs into your forum post. If you must use AI’s words, keep it to a short paragraph, or less. And, always be sure to indicate that the text was AI generated. Better yet, use AI to track down original sources and cite those directly and summarize what you’ve learned using your own words.”

@AbramBaileyAuD said:

Perhaps the rule could be something like: “Don’t paste long AI outputs into your forum post. If you must use AI’s words, keep it to a short paragraph, or less

Exactly, this is part of the problem, the other is when it’s clearly incorrect, and yes to cite original sources, it’s always good to be able cross reference, which a basic Google search is good for anyway.

Edited

Since is probabilistic, it’s guaranteed to be wrong sometimes. It’s definitely been wrong on . The modified rule (don’t just copy and paste long responses) seems like a great idea.

I’d like to see a requirement for decent word usage, or at least proper use of words that sound like ‘yore’. ‘Your’ and ‘you’re’ are not the same word.

Not that I’m a grumpy old man or something….

In reply to @AbramBaileyAuD

@AbramBaileyAuD I’d prefer a stricter ‘no ’ rule.

Even if forum users shorten the AI output and flag it as such, I don’t see the benefit. It is clear that AI output is unreliable at best. Why do we need it here?

Right now, every time I see someone mention AI or ChatGPT/etc., I scroll right past because I don’t want to waste my time on something that may or may not be true.

Anybody who wants to use AI privately can do so. It doesn’t have to be posted here.

In reply to @AbramBaileyAuD

@AbramBaileyAuD said:

I don’t know that I want to be that black and white about AI use.

The ‘problem’ is that while these threads might not contain absolute ‘truths’, they do offer a fairly balanced source of facts and personal experience. At the minute (from a few weeks ago), this source isn’t corrupted by errors and extrapolation.

If you allow AI content in, you create the risk of positive feedback loops, due to the volume of bilge/data from the AI, the false information gets propagated as legitimate source material, thereby destroying the integrity of the site as a source of information. It happens with any media to a sense, but the loop is much tighter with web based AI.

The wider question is whether you want to predicate the demise of the integrity of the forum by allowing the ‘machine’ in.

I guess point me to the posts that are concerning so I can review the situation. I don’t really see the forum on a slippery slope to overwhelming amount of generated content… but, I’m willing to have a deeper look.

For a lot of my searches, this site comes up in a dominant role in having the answers. I agree with the feedback concern. If the answer someone gives here from is bogus, that will feed into further AI chaos. And I find a lot of answers from AI incorrect. A lot of exceptions to generalities are missed. Generational distinctives in platform changes are often missed. For some areas, it isn’t even worth asking AI.

WH

In reply to @AbramBaileyAuD

Upcoming Review on Audiologist who does a sophisticated battery of tests beyond checkbox
Initially I was not going to respond as you are not the first person on this forum or first audiologist who has said to me that people with a hearing loss must manage their expectations! However, I have decided to make a reasoned case for why I no longer believe that this is acceptable. Note I also used to believe the same as you that I had to be realistic about the fact that a hearing aid could never replace a human ear. About me in a business context I used to run a small business. One val…

This one went down the rabbit hole.

In reply to @AbramBaileyAuD

@AbramBaileyAuD, I think the forum is still fine right now, but there is an uptick in people posting output.

I think the question to ask is whether any of the user posts containing AI output have actually contributed anything. Did AI offer valid solutions? New insights? I think in too many cases the answer has been ‘no’.

It’s one thing for someone asks a question like: “I checked with ChatGPT and tells me ‘xyz’. Is ‘xyz’ correct”. That’s OK.

But too often people post AI output as a response to help other users, and then it takes extra effort from actual humans to debunk.

My own experience is that I get really good information from GPT 5.2 extended thinking and the paid version of Gemini… free models are generally really bad right now, and tend to oversimplify and hallucinate. But, I’m seeing the situation improve in general when it comes to the best paid models. And who knows what the future holds as these things continue to improve. I suppose we could put a full ban in place until more people are using reliable models, but there is also a counter risk that people are not able to post legitimate content that would be helpful. I have run into this myself on reddit when I see a bunch of misinformation in a thread and want to quickly provide some better information, but don’t have the time to write it myself… I query AI, get decent information (I know because I am an expert in a couple of fields, like audiology and web development) but cannot share it due to AI restrictions. I know this seems like a made up situation, but I can assure you, it is not… So, just something to think about… this is a fluid situation, and I’m open to picking a side right now, but I wouldn’t want to guarantee I’d never reconsider… @Um_bongo regarding that example post you shared… I think that would one would have been ok with the rule I proposed above.

Active discussions

What the forum has been talking about this week.