Should a member be allowed to make a post which is AI generated or AI mixed without disclosing such use of AI as part of the post?

[tima]

Once again I find it tough to understand that AI is used for spelling and grammar checks. Whenever I write a post any word which is misspelled is underlined in red and the properly spelled word is shown. So also is the same with grammar. In fact our posts now are so intuitive that the sentence is often finished for us and more often is correct. Let’s get beyond using AI to spell check because IMHO that’s bull crap. Now the word polishing has surfaced. To me this is AI manipulation plain and simple Call a spade a spade Ron What you’re saying is doing nothing except fostering a mistruth. Really someone needs to use AI to spell check. I call BS on that. And as far as polishing what a dumbass word. To me this is someone who is asking AI to make him sound like an expert by AI manipulation . If anybody disagrees go for it and make your case. To me you are burying the intention to deceive by calling it spell check and polishing. I say give me a break and call it what it is

Yes, exactly. I support this view. This 'polishing' is a weasel word that leads to the slippery slope of 'how much AI is acceptable'.

You put your own text into the thing and you decide if you want just spelling and grammar checked, or if you also want also it kind of polished linguistically. I don't know how else to put it.

Too vague. I know how to put it: it is AI generated content because it is no longer the author's words. Don't put your "text into the thing" and the issue goes away.

This attempted bifurcation between "AI generated content" and "AI polishing" is more needless complexity that will lead to more contention and denigration of forum integrity.

 

[tima]

We should take this somewhere else…

Please do. Just because Denney The Third pays to be here doesn't mean he's not disruptive.

 

[Al M.]

When Open AI released an AI watermarking tool, someone applied it to the Bible, which it then claimed was AI generated.

Thanks. Kinda makes my point, doesn't it?

See also here:

https://cybernews.com/editorial/your-essay-was-ai-generated-so-was-the-bible-harry-potter/

 

[microstrip]

1. Tarnish Impact:
For audio frequencies (20 Hz - 20 kHz), the effect of tarnish on a 0.25 mm silver wire would be negligible. The tarnish layer, primarily silver sulfide, is still conductive, albeit less so than pure silver. The thin tarnish layer would not significantly affect the wire's overall conductivity or the signal transmission in the audio range.

2. Skin Effect:
At audio frequencies, the skin effect is minimal. Even if tarnish slightly alters the wire's surface conductivity, it wouldn't significantly impact signal propagation for audio frequencies.

3. Air vs. PTFE Dielectric:
The air dielectric, despite allowing for some tarnish formation, would still be preferable to PTFE for several reasons:

- Lower Dielectric Constant: Air has a dielectric constant very close to 1, while PTFE is around 2.1. This means less signal alteration and energy storage in the dielectric.
- Less Signal Smearing: Air introduces far less time smearing and phase distortion compared to solid dielectrics like PTFE.
- No Heat-Induced Tarnish: Unlike PTFE application, which can cause heat-induced tarnish during manufacturing, air dielectric allows the wire to remain in its original state.

4. Parallel Wire Configuration:
The presence of the larger 1 mm wire in parallel would further mitigate any minor effects of tarnish on the 0.25 mm wire, as the larger wire would carry most of the current due to its lower resistance.
View attachment 145535

In conclusion, for audio frequency transmission, the slight tarnish on a 0.25 mm silver wire in an air dielectric would be far less detrimental than the combined effects of PTFE dielectric properties and potential heat-induced tarnish from PTFE application. The air dielectric configuration results in superior signal transmission characteristics for audio applications, even with some surface tarnish present.

Your argumentation is interesting. You refer to physics that proves that effects you do not care can be ignored in the audio bandwidth and then refer to effects that the same kind of physics say that are not relevant in the audio bandwidth.

Dielectric constant and parallel wire configuration have no relevance in the audio bandwidth. Can you tell us what is exactly signal smearing? Are you addressing triboelectricity and cable microphony?

 

[microstrip]

You’re not thinking about what signal transfer actually is, the wire is a wave guide for an electromagnetic field. In this regard tarnish has virtually no effect whatsoever on signal propagation. As for separating cables, the carbon fiber spacer discs for maintaining the air dielectric keeps the pure silver conductors precisely spaced in air and ensures consistent electromagnetic field propagation, avoiding the dialect absorption and capacitance issues of solid insulation keeping the pure silver conductors precisely spaced, and avoiding the dielectric absorption and capacitance issues of solid insulation. Furthermore, the use of carbon fiber provides a passive RF dampening for cleaning up the audio signal from potential high frequency interference.

As I said earlier, we’ve listened to cables up to five years old, any differences due to tarnish are virtually imperceptible. It’s the kind of thing maybe you think you hear a difference, maybe you don’t. However that exact same cable, where the silver conductors are encompassed in a PTFE or Teflon dielectric, sounds significantly lower resolution due to switching from air to Teflon. You seem like the sort of guy, who can engage endlessly beating a dead horse. I’ve made my points here, and I’m not going to waste time with circular banter. Have a good day.

Thanks for avoiding answering my post and posting again your usual advertisement slogans. Are you wanting to prove us that you speaking to your iPhone can be more dangerous than AI?

 

[Rensselaer]

This thread asks if we should allow AI to be used (if so, to what degree) on this forum.

Before deciding, we must at least have determined what the majority of the subscribers want from this forum, perhaps another polling.

I for one, seek to learn from the experience-led perspectives of the other subscribers. A place where they that have experience with the very finest examples of our obsession available can share their experience so that we all gain insight (and no-doubt save money).

AI (to date) scours everything on the internet then draws from it the information one wishes it to find. Meaning, it will look at articles that may have been fictitiously created to sell products, and from them construct an elaborate and seemingly valid argument for or against, as requested.

With how much certainty can we accept then what AI provides? Let me give an example. I asked my search engine a simple question, who makes the best-sounding amplifier? This was my answer:
Who makes the best-sounding amplifier

Although that question would have caused quite an argument amongst our members, I seriously doubt that any would have cited even one of the amplifiers listed on that web page. What that tells me is that information on the web can not be trusted to be the absolute truth regardless of subject. I mentioned on another thread a story I read about lawyers being struck off after using AI to construct their legal argument. Apparently, all the case law referenced by AI was made up, totally fake. AI is in its' infancy and like children it wants to please you; it just hasn't the maturity to know right from wrong, truth from fiction.

There are other members whom, it has been alleged, are using AI to bolster their arguments as to the superiority of their products for the purpose of selling their wares to the viewers (A fair amount of too and from them is what led to this thread being started). It would thus seem that merchants who use this forum as (free?) advertising would no-doubt want to use AI (to create convincing argument, at a very low price). I would also point out that the arguments between those vendors (which snake oil formula is best) doesn’t harm business, quite the opposite (wasn’t it P.T. Barnum who said there is no such thing as bad publicity).

So, we need to decide as a group whether we want to be in a forum of highly experienced audiophiles sharing ideas accrued from using the finest equipment ever created for reproducing music in our homes, or members of a target audience (with the means to afford) for venders of very expensive products that use (probably false, certainly deceptive) AI to convince us to part with our hard earned dollars and buy items that may make a “difference”, not necessarily “an improvement”.

Let the poll begin.

 

[facten]

I'll vote for a forum of highly experienced audiophiles sharing ideas

 

[Iamrael]

Would it be a viable option or consideration to use the level or status of any given member (“Paid Advertiser”, “Expert”, “Dealer”, etc.) as a determining factor in requiring either full disclosure or banning the use of AI all together?

(……Maybe even just as a starting point and expanding as things work through)

This wasn’t a rhetorical question btw!
On another note, perhaps allow all post, whether deemed to be “polished” or not, but put a “score” under the members name based on either the average AI scores of all the members’ posts, just the one…or both avg. and latest post.

 

[microstrip]

(...)
Let the poll begin.

There is still a point it is not clear to me. As far as I see it (I may be wrong) , independently or our wishes, WBF will not have AI detection before XenForo, who provides the software package that runs the forum, implements such plug-in. Is there any date scheduled for such launch? Are we debating mostly for the enlightened of a nice discussion?

 

[wil]

I like the idea that the Monica thing incorporates several AI detectors to yield a composite result. I personally would find this more authoritative and trustworthy than any particular single AI detector.

I have a solution. In the WBFTOA, require that any and all AI postings be done in Olde English.

 

[Rensselaer]

If I understand what you are asking, Microstrip, I would think that the inclusion of AI detection software would not be necessary if the majority of members vote to allow merchants to use this forum to peddle their wares using AI.

 

[treitz3]

I'll vote for a forum of highly experienced audiophiles sharing ideas

Ah, you mean the main goal of the forum? The very first thing one reads before reading the Terms of Service...

Our goal is to have a friendly place where everyone shares ideas and information without the level of bickering and angst that other forums tend to create.

Tom

 

[Rensselaer]

Ah, you mean the main goal of the forum? The very first thing one reads before reading the Terms of Service...

Our goal is to have a friendly place where everyone shares ideas and information without the level of bickering and angst that other forums tend to create.

Tom

So you never added: “ and to provide manufacturers with a target audience to direct their AI-created advertising to” ?

 

[Steve williams]

You’re not thinking about what signal transfer actually is, the wire is a wave guide for an electromagnetic field. In this regard tarnish has virtually no effect whatsoever on signal propagation. As for separating cables, the carbon fiber spacer discs for maintaining the air dielectric keeps the pure silver conductors precisely spaced in air and ensures consistent electromagnetic field propagation, avoiding the dialect absorption and capacitance issues of solid insulation keeping the pure silver conductors precisely spaced, and avoiding the dielectric absorption and capacitance issues of solid insulation. Furthermore, the use of carbon fiber provides a passive RF dampening for cleaning up the audio signal from potential high frequency interference.

As I said earlier, we’ve listened to cables up to five years old, any differences due to tarnish are virtually imperceptible. It’s the kind of thing maybe you think you hear a difference, maybe you don’t. However that exact same cable, where the silver conductors are encompassed in a PTFE or Teflon dielectric, sounds significantly lower resolution due to switching from air to Teflon. You seem like the sort of guy, who can engage endlessly beating a dead horse. I’ve made my points here, and I’m not going to waste time with circular banter. Have a good day.

Just for the records as I know microstrip is modest and at many times can be contentious but for the record IIRC he is a professor of electrical engineering at the University of Lisbon.

 

[Steve williams]

There is still a point it is not clear to me. As far as I see it (I may be wrong) , independently or our wishes, WBF will not have AI detection before XenForo, who provides the software package that runs the forum, implements such plug-in. Is there any date scheduled for such launch? Are we debating mostly for the enlightened of a nice discussion?

There is nothing in the works as I have mentioned before at Xenforo to suggest they are working on their own AI detector . What I did mention is that on their own forum for users of their platform they are selling 3rd party AI detectors

 

[wil]

AI generated content is one of the things this thread is about.

Even in the draft Term of Service I distinguish very conscientiously between two different uses of AI:

posts which are generated by AI

and

posts which are member drafted and then processed through AI

These are two separate things; two separate uses of AI.

I think we're all agreed that nobody wants AI generated content where a member asks AI a question and the member just posts the answer as his/her own. That's not kosher.

The harder question is when AI is used at what I am calling the back-end, for spelling correction and grammar correction and, possibly, polishing. Depending on the application and how the user uses the application the polishing can be dialed up and down (I think). I'm being a little figurative here, but if the polishing is very strong, the originally written text may become so polished as to flunk the AI detector. But if it's just used for spelling and grammar and the polishing is at a low setting, so to speak, then the post will sound natural and will pass the AI detector.

I don’t want to read AI polished prose. And it’s a slippery slope.

 

[Ted Denney III]

I have a solution. In the WBFTOA, require that any and all AI postings be done in Olde English.

Hark! I have a remedy most fitting. In the hallowed halls of WBFTOA, let it be decreed that any and all missives penned by artificial minds be writ in the tongue of our forebears, the noble Olde English.

 

[DaveC]

Now in my own words:


Again, AI is not a magic eight ball; it won’t make an idiot intelligent, but it will make a highly intelligent person a genius, and a genius an Albert Einstein or Nikola Tesla.

1000% No

I guarantee it's not helping you at all, Ted.

 

[Ron Resnick]

AI . . . is a tool that you can use to . . . better communicate your ideas to a target audience.

I agree. But this is a separate point.

What is coming through to me loud and clear on this thread is that we hobbyists do not want to read on our hobby forum what AI is producing. We would rather stay with authentic -- if less than perfect -- spelling, grammar and syntax. We do not want polished or perfect-sounding or robot-sounding posts. We do not want posts to read as homogenized gradually over time. It is as simple as that.

For the record, I am using AI now to radically accelerate the evolution of Synergistic Research technologies I pioneered over the past 30 years. Now, when a new product is launched, it’s not a next-gen evolution; it’s five or six generations advanced because we’re able to leapfrog from one generation to the next, extrapolating our data, testing against art in non-related fields that may correlate to what we’ve been doing in-house, and ultimately, dramatically moving the ball forward in terms of our products’ performance. Ultimately, all of this serves one purpose: to improve subjective outcomes for our products—the most human thing possible.

I think it is great that you are finding AI useful in your professional life. What we hobbyists want to read here is a totally different question.

It is very simple to avoid friction over AI with other members. Do not run your draft WBF posts through AI before posting. We would rather see spelling mistakes and imperfect grammar.

Members-only clubs have rules. No AI is WBF informal culture. This thread discusses whether this culture should be elevated to a formal rule.

 

[Rensselaer]

Was the above at least partially written by AI and then modified to avoid detection?


First An AI Generated Version of the Text

Second A Walk Through of Alterations That Could Transform It into the More Casual Version (Potentially Evading AI-Detection Heuristics)

1. AI Generated Version

Subject: Evaluating the Use of Artificial Intelligence on Our Forum

This discussion thread is dedicated to assessing whether artificial intelligence should be permitted on this forum—and if so, to what extent. Before any policy decisions are made, it is advisable to conduct a poll to ascertain the preferences of the majority of subscribers.

The forums primary objective is to facilitate the exchange of experiential knowledge among audiophiles who use top‑tier equipment for music reproduction. Members are encouraged to share insights drawn from their extensive hands‑on experience, thereby enriching the collective understanding and potentially reducing individual costs.

Currently, AI systems methodically search the internet, retrieving and synthesizing information—including content that may serve promotional purposes. For example, when asked a question like “Who makes the best-sounding amplifier?â€, an AI might compile data from a range of sources and produce an answer that appears both detailed and valid. This capability raises questions about the certainty with which we can accept AI‑generated responses.

Additionally, there are reports suggesting that certain forum participants may be using AI to strengthen their arguments regarding product superiority. Such usage could provide a cost‑effective means of advertising, as vendors might leverage AI‑assisted content creation to promote their products. This dynamic calls into question whether the forum should serve as a venue for genuine, experience‑based discussion or evolve into a platform influenced by commercially motivated, AI generated arguments.

We now invite all members to participate in a poll so that the community can decide its preferred direction moving forward.

2. Alteration Process: From AI Generated Formality to a More Casual, Humanized Version

Below is an explanation of several straightforward modifications that can transform the above AI‑style text into a version like the one you originally provided alterations that, in effect, might help the text or evade typical AI-detection heuristics by introducing a more idiosyncratic, conversational tone:


1. Tone and Voice Adjustment


Formal style is advisable to conduct a poll to ascertain the preferences

Altered Version: Before deciding, we must at least have determined what the majority of the subscribers want perhaps another polling.


Result: Shifts from formal language to a more conversational, informal tone.

2. Personalization and Subjectivity

Formal: Members are encouraged to share insights drawn from their extensive hands ‘on experience.

Altered: I for one, seek to learn from the experience-led perspectives of the other subscribers. A place where those who have used the very finest examples of our obsession can share their experience so that we all gain insight (and no-doubt save money).

Result: Injects personal opinion and colloquial expressions that are less typical of formal AI output.

3. Colloquial Phrasing and Idiomatic Expressions

Formal: AI systems methodically search the internet, retrieving and synthesizing information

Altered: AI (to date so far) scours everything on the internet and draws from it the information one wishes it to find.â€

Result: Introduces parenthetical asides and informal phrasing that lend a human, conversational quality.

4. Inclusion of Rhetorical Questions and Examples

Formal: For example, when asked a question like ‘Who makes the best-sounding amplifier? an AI might compile data

Altered: Let me give an example. I asked my search engine a simple question, Who makes the best-sounding amplifier? This was my answer: Who makes the best-sounding amplifier

Result: The altered version uses a first person anecdote and a direct, less polished narrative style.


5. Critique and Humor in the Discussion

Formal: There are reports suggesting that certain forum participants may be using AI to strengthen their arguments

Altered: There are other members whom, it has been alleged, are using AI to bolster their arguments as to the superiority of their products for the purpose of selling their wares

Result: The informal version adds humor and a slightly irreverent tone (including parenthetical commentary on free advertising and historical quips like the P.T. Barnum reference) that contrasts with the neutral style of the generated text.


6. Overall Structural Looseness

The formal version follows a strict, logical structure with clear demarcations.

The altered version intersperses commentary, informal asides, and occasional grammatical looseness the arguments between those vendors doesn't harm business that reflect a more spontaneous, human generated discussion.

Result: Such structural looseness is often less common in AI generated texts and may help the altered version evade detection algorithms that look for overly consistent syntax and formality.

By applying these modifications introducing colloquial language, personal commentary, informal punctuation, and a less rigid structure an initially AI generated draft can be altered into a version that exhibits characteristics more typical of human-authored text. This process, while straightforward, illustrates how minor stylistic shifts can affect the perceived origin of a text, potentially complicating AI-detection efforts.

The above is a deep dive using AI to see if the quoted text may have started as an AI-prompted argument that was then humanized to avoid AI detection. When analyzed, it showed a moderate probability of being at least partially written by AI and then disguised. All of this, however, completely misses the point.

You (IMHO) are quite full of it.