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?

[italia]

What's your *proof* of AI generated text with respect to this manufacturer? I am not asking for suspicions, anyone can have these, but for *proof*. The "judgment" of an AI "AI detector"?

Not that I particularly like the writing style myself, but dislikes and differences in taste are not "proof".

I'm just going to leave this here. The lack of consistent/ reliable efficacy of AI detection gives just about anyone plausible deniability.

"GPTZero advertises itself as producing the “most precise, reliable AI detection results on the market.” ZeroGPT bills itself as “the most Advanced and Reliable.” And Winston AI claims “unmatched accuracy” with a “99.98% accuracy rate.”

While researchers found that most private companies studied had calibrated their models with sensible false-positive rates, the same wasn’t true for many tools that use open-source models.

Callison-Burch said the trouble with accuracy rates is that they often neglect false positives: Anyone can catch 100% of AI-generated content if they’re willing also to flag all or most human-generated content as being AI-generated, for instance. He and his team found that adjusting models’ false-positive rates to what he called a “reasonable” level greatly reduced the ability of the models to detect AI-generated content.

“These claims of accuracy are not particularly relevant by themselves,” he said."

Source: https://edscoop.com/ai-detectors-are-easily-fooled-researchers-find/

 

[Ron Resnick]

Here for your consideration is a new draft of a possible new Term of Service prohibition on AI posts. Presently it seems to me that the AI detector controversy is too much of a morass to subject members to the risk of suspensions and bans on the basis of false positives.

Once Xenforo has a built-in AI detector the moderators can flag suspected AI posts, and will discuss such posts with the members who posted them, showing such members the evaluation by the AI detector that the post is probably 100% AI. This process should be transparent to the members.

The moderators will have the ability to delete 100% probability AI posts. The moderators will have the ability to informally warn members who the moderators believe have made AI posts. But I don't think AI detectors or member confidence in this overall process is mature enough to subject members to suspensions and bans.


NEW PROPOSED DRAFT OF AI TERM OF SERVICE

Posts which are generated by artificial intelligence applications (“AI”), and posts which are member drafted and then processed through AI (collectively, "AI Posts"), are prohibited.

WBF utilizes the AI detection application built into Xenforo to detect AI Posts. An AI Post which is proven transparently to an AI Post accused member to be 100% probability AI may be deleted by forum administrators. The posting of AI Posts shall not be subject to suspensions or bans.

 

[microstrip]

What's your *proof* of AI generated text with respect to this manufacturer? I am not asking for suspicions, anyone can have these, but for *proof*. The "judgment" of an AI "AI detector"?

No one presented any proof of nothing in this debate. I checked parts of it it on several AI detectors and got between 93 and 100%.Two other posters in other treads did the same. Readers and forum owners will decide for themselves.

 

[PYP]

I'm just thankful WBF does use a bullsh^t detector and auto-delete. Many of my posts would be gone. But so would some threads.

 

[Ron Resnick]

Ted weren t you the first caught with AI on WBF?
Causing all this ....

There's a famous expression that "bad facts make bad law."

The analogy here is that a one-off problem that causes a kerfuffle might make a bad Term of Service. Let's not let that happen.

 

[Ron Resnick]

As they say The dogs bark, but the caravan moves on

I had to ask ChatGPT what this means.

 

[Al M.]

NEW PROPOSED DRAFT OF AI TERM OF SERVICE

Posts which are generated by artificial intelligence applications (“AI”), and posts which are member drafted and then processed through AI (collectively, "AI Posts"), are prohibited.

WBF utilizes the AI detection application built into Xenforo to detect AI Posts. AI Posts may be deleted by forum administrators. The posting of AI Posts shall not be subject to suspensions or bans.

Again, how do you reliably determine that something is AI generated?

There's a famous expression that "bad facts make bad law."

The analogy here is that a one-off problem that causes a kerfuffle might make a bad Term of Service. Let's not let that happen.

Exactly.

 

[PYP]

I had to ask ChatGPT what this means.

well, you can still share the info until the new TOS apply.

 

[WBF Team]

I found that XenForo has several AI based plug-ins to help forum moderation and detection of spam or inappropriate content. Are they currently in use?

No.

 

[Ted Denney III]

When the debate becomes not about the content of something that was said, but rather the tone in which it was expressed, and once the capricious label of AI is applied, it can be used as a censorship tool.

Case in point: after being attacked by a forum member that turned out to be a competitor I had never heard of, I took umbrage and responded in kind. This exchange heated up, and a ridiculous claim was made that encompassing Teflon dielectric is superior to air because pure silver, when oxidized, becomes a poor conductor. This is a false argument on so many levels, and I dismantled it as such. Instead of my response being replied to, instead of the points I made being refuted (which they could not be), it was asserted that my response was written with AI.

The point of free speech is more speech. Points should be analyzed and debated, not attacked because it is presumed that they may have been made either with the help of or exclusively through AI. People who have a history of writing ad copy for many decades, or who are well-versed in communicating complex ideas, will always come across as “AI.”

Better to address the content of points made than to attack the style of delivery.

The above was verbally dictated to my iPhone, and then corrected for punctuation and grammar.

 

[Ron Resnick]

I'm just going to leave this here. The lack of consistent/ reliable efficacy of AI detection gives just about anyone plausible deniability.

What about Monica which employees GPTzero and two or three other AI detectors on a composite analysis basis?

 

[Ron Resnick]

Again, how do you reliably determine that something is AI generated?

Take a look at this Monica thing.

 

[thedudeabides]

I never get tired of winning

I don't know about winning if that means selling product based on your descriptions. Your writing style is extremely hyperbolic. You sound like a used car salesman on steroids. That alone would discourage me from seriously considering purchase.

I like Conrad Johnson's marketing moto. It just sounds right.

 

[DaveC]

When the debate becomes not about the content of something that was said, but rather the tone in which it was expressed, and once the capricious label of AI is applied, it can be used as a censorship tool.

Case in point: after being attacked by a forum member that turned out to be a competitor I had never heard of, I took umbrage and responded in kind. This exchange heated up, and a ridiculous claim was made that encompassing Teflon dielectric is superior to air because pure silver, when oxidized, becomes a poor conductor. This is a false argument on so many levels, and I dismantled it as such. Instead of my response being replied to, instead of the points I made being refuted (which they could not be), it was asserted that my response was written with AI.

The point of free speech is more speech. Points should be analyzed and debated, not attacked because it is presumed that they may have been made either with the help of or exclusively through AI. People who have a history of writing ad copy for many decades, or who are well-versed in communicating complex ideas, will always come across as “AI.”

Better to address the content of points made than to attack the style of delivery.

The above was verbally dictated to my iPhone, and then corrected for punctuation and grammar.

No Ted, you just proved that you don't understand how silver corrodes, yet you offer cables with an air dielectric anyways. This is called incompetence. You are not qualified or competent to design and build the products you sell. But it goes further than that. You are selling a product that will not last because it will corrode, which is unethical.

You still insist it's "oxidized" but apparently you don't even know what that means, because it's NOT oxidized. Silver tarnish is a sulphide, not an oxide. For someone who sells cables that use wire that isn't protected from corrosion, maybe you should learn how silver corrodes? If you did, you might not choose to sell a product that will everntually need to be thrown in the trash due to corrosion.

I could easily rebuke the flawed logic your AI model used, but I refuse to argue with AI.

This is AI, and nobody needs an AI detector to understand that you didn't write this post:

Ted Denney III said:
Thank you for sharing your perspective. However, I must respectfully disagree with several points in your response, as they contain inaccuracies and unfounded personal attacks.

Firstly, your focus on the specific composition of silver tarnish misses the broader point about dielectric properties. While you’re correct that silver sulfide is more common than silver oxide in everyday environments, this distinction doesn’t invalidate the core argument about air versus Teflon dielectrics.

Your claim that I lack the technical background for cable design is both unfounded and irrelevant to the discussion at hand. It’s a classic ad hominem fallacy that attempts to discredit the argument by attacking the person rather than addressing the actual points raised.

The 2001 forum post you’ve cited, while interesting, is not a credible scientific source. It’s an informal, outdated discussion that doesn’t reflect current research or advancements in audio technology over the past two decades.

Your statement about my competence is not only unnecessarily personal but also demonstrates a logical fallacy known as the argument from authority. My points stand or fall on their own merits, not on your perception of my qualifications.
The core argument about air versus Teflon dielectrics remains valid and is based on well-established principles of electromagnetic theory. The lower dielectric constant of air compared to Teflon can indeed lead to less signal alteration in certain circumstances.

I’d encourage you to engage with the actual technical points raised, rather than resorting to personal attacks or outdated sources. A constructive dialogue based on current research and empirical evidence would be far more beneficial to the audio community.

Let’s focus on advancing our understanding of cable design through respectful, fact-based discussions. This approach serves everyone better and contributes to the collective knowledge in our field.

 

[Ron Resnick]

When the debate becomes not about the content of something that was said, but rather the tone in which it was expressed

I agree in theory, but in practice I do not want to read posts which sound inauthentic and overly polished and read like they were written by a robot.

Posts should be written in members' natural style, grammar mistakes and all, without excessive AI polishing. This is supposed to be a fun hobby forum, not 11th grade English class.

WBF thrives on real discussions. Posts should reflect genuine opinions and experiences, not AI-polished rewrites.

and once the capricious label of AI is applied, it can be used as a censorship tool.

I agree this is a very big risk.

The point of free speech is more speech. Points should be analyzed and debated, not attacked because it is presumed that they may have been made either with the help of or exclusively through AI. People who have a history of writing ad copy for many decades, or who are well-versed in communicating complex ideas, will always come across as “AI.”

Better to address the content of points made than to attack the style of delivery.

The above was verbally dictated to my iPhone, and then corrected for punctuation and grammar.

I ran your entire post through Originality.ai. The result was: "We are 57% confident that text is AI-generated."

Since the AI Post threshold I built into the draft proposed text of the AI Term of Service is 100%, your correction for punctuation and grammar was not a problem.

 

[andromedaaudio]

Can somebody please ask AI whats the best cable ?

 

[Ted Denney III]

When the debate becomes not about the content of something that was said, but rather the tone in which it was expressed, and once the capricious label of AI is applied, it can be used as a censorship tool.

Case in point: after being attacked by a forum member that turned out to be a competitor I had never heard of, I took umbrage and responded in kind. This exchange heated up, and a ridiculous claim was made that encompassing Teflon dielectric is superior to air because pure silver, when oxidized, becomes a poor conductor. This is a false argument on so many levels, and I dismantled it as such. Instead of my response being replied to, instead of the points I made being refuted (which they could not be), it was asserted that my response was written with AI.

The point of free speech is more speech. Points should be analyzed and debated, not attacked because it is presumed that they may have been made either with the help of or exclusively through AI. People who have a history of writing ad copy for many decades, or who are well-versed in communicating complex ideas, will always come across as “AI.”

Better to address the content of points made than to attack the style of delivery.

The above was verbally dictated to my iPhone, and then corrected for punctuation and grammar.

No Ted, you just proved that you don't understand how silver corrodes, yet you offer cables with an air dielectric anyways. This is called incompetence. You are not qualified or competent to design and build the products you sell. But it goes further than that. You are selling a product that will not last because it will corrode, which is unethical.

I could easily rebuke the flawed logic your AI model used, but I refuse to argue with AI.

This is AI, and nobody needs an AI detector to understand that you didn't write this post:

Ted Denney III said:
Thank you for sharing your perspective. However, I must respectfully disagree with several points in your response, as they contain inaccuracies and unfounded personal attacks.

Firstly, your focus on the specific composition of silver tarnish misses the broader point about dielectric properties. While you’re correct that silver sulfide is more common than silver oxide in everyday environments, this distinction doesn’t invalidate the core argument about air versus Teflon dielectrics.

Your claim that I lack the technical background for cable design is both unfounded and irrelevant to the discussion at hand. It’s a classic ad hominem fallacy that attempts to discredit the argument by attacking the person rather than addressing the actual points raised.

The 2001 forum post you’ve cited, while interesting, is not a credible scientific source. It’s an informal, outdated discussion that doesn’t reflect current research or advancements in audio technology over the past two decades.

Your statement about my competence is not only unnecessarily personal but also demonstrates a logical fallacy known as the argument from authority. My points stand or fall on their own merits, not on your perception of my qualifications.
The core argument about air versus Teflon dielectrics remains valid and is based on well-established principles of electromagnetic theory. The lower dielectric constant of air compared to Teflon can indeed lead to less signal alteration in certain circumstances.

I’d encourage you to engage with the actual technical points raised, rather than resorting to personal attacks or outdated sources. A constructive dialogue based on current research and empirical evidence would be far more beneficial to the audio community.

Let’s focus on advancing our understanding of cable design through respectful, fact-based discussions. This approach serves everyone better and contributes to the collective knowledge in our field.

Like I said, I’ve never heard of you. Never ever, and I’ve been doing this since I was in my 20s, and I’m now 60. And you are completely wrong that an all-encompassing PTFE dielectric around a solid silver conductor is preferable to air. The incursion that the PTFE creates on the electromagnetic transfer—it’s not electrons running down a wire, Dave; it’s the EM field around the wire. That’s what signal is, Dave. And when you put a solid dielectric around the conductor, that impedes the electromagnetic transfer more so than do various types of silver oxide in an air dielectric. In other words, the deleterious effects of PTFE outweigh any negative consequences of silver when it tarnishes in an air dielectric.

The above was verbally dictated to my iPhone, and corrected for punctuation and grammar.

 

[DaveC]

When you encounter posts such as Ted's that are very obviously AI, a detector isn't required.

I also think it's not required for almost every single member here. But Ted has demonstrated he's willing to lie about it. I think he's a group of ONE. He's the reason this thread exists, and his denial of his obvious use of AI is problematic.

 

[DaveC]

Like I said, I’ve never heard of you. Never ever, and I’ve been doing this since I was in my 20s, and I’m now 60. And you are completely wrong that an all-encompassing PTFE dielectric around a solid silver conductor is preferable to air. The incursion that the PTFE creates on the electromagnetic transfer—it’s not electrons running down a wire, Dave; it’s the EM field around the wire. That’s what signal is, Dave. And when you put a solid dielectric around the conductor, that impedes the electromagnetic transfer more so than do various types of silver oxide in an air dielectric. In other words, the deleterious effects of PTFE outweigh any negative consequences of silver when it tarnishes in an air dielectric.

The above was verbally dictated to my iPhone, and corrected for punctuation and grammar.

Ted, what are your actual credentials? You have a very basic, elementary understanding of electrical phenomenon that makes me believe you have no real education or knowledge of the subject.

 

[italia]

I like Conrad Johnson's marketing moto. It just sounds right.

Is Lew still around? I recall meeting him when I worked for a dealer many decades ago. Seems like his McCormack Audio thing didn't survive after he left C-J. So many interesting people in the industry back in the day