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?

[andromedaaudio]

Can anyone write code for a digital BOT and one for a Analogue BOT .
Or China AI against Silicon valley AI

Then have a separate thread where the 2 can fight it out at digital versus analogue , that would be cool


Just make a seperate AI section on WBF

 

[andromedaaudio]

Rumour goes the only thread not affected by AI yet is Peters " Natural sound thread

Ps did anyone do the AI check on Hifi publications ?

 

[microstrip]

I started this thread as everyone knows. I want to have the feeling that everything I read by someone posting is legitimate and not from some bot. Surely we can still think, question and understand and come to conclusions. I don't want anyone to gussy up their posts with anything other than their own thoughts. Nor do I want to read posts that are designed to pump up one's desires to be construed as an expert.

Ron, with all due respect the poll speaks primarily for the most vociferous of posters on WBF. It's just like Mike Lavigne said. We have over 30,000 members but the reality is there is a core group of members who post regularly and IMO are the "brains" behind WBF and need to be heard. Frankly I am not surprised by the results of the poll so far.

Having said this I too feel very strongly that we do whatever we can to prevent any member from using AI to enhance his posts without full disclosure at the time of posting. I don't want to be duped nor do I want someone pumping his chest with each post that he is the sole expert and the rest of us just better listen to what he has to say

To that extent I feel that the voting majority are the core group of members here on a daily basis and posting the most and therein is the crux of this debate. I feel that we as owners now have an obligation to add to our Terms of Service a clause which makes members aware that it is not allowed on WBF. I also believe that virtually 100% of our regular posters will never use it and if they do it will be with full disclosure. There will be new members who join who might not be aware of such a rule and therefore a soft (informal) warning by the mods would iMO be sufficient enough that this member will never post with AI again and/or without full disclosure. No one is out to ban anyone but we are all wanting a level playing field and that is not too much to ask of any of us,

Here is what I propose be added to our TOS.....I believe that the majority of members are strongly opposed to the use of AI in any form .

I propose "members shall not use AI in any fashion unless such is divulged in the post. Failure to do such will result in removal of the post and an INFORMAL warning issued. If further use of AI by the same member continues, the post(s) will be removed and a FORMAL warning will be issued. Pursuant to our use of a ladder system of warnings. repeated violations will result in a temporary ban and if continues a permanent ban will result.

It is my feeling that if a member does this unknowingly of the TOS and an INFORMAL warning issued, my bet is they will comply and never do it again. I trust our members as good people and an. INFORMAL warning is meant to serve such purpose. We have used a ladder system for years and it works perfectly . If the member continues to be an offender, the ladder system gives him ample opportunity to weigh his future posts inasmuch as the ladder system after several formal warnings will result in a temporary ban and if said person continues to flick us the finger and persist in such posts a permanent ban will result

Our mods are completely fair and our TOS is their constitution. No AL detector is fool proof and I believe that this thread alone will give members pause should they think of using AI generated posts so that in the long term I really believe this issue at least for now will be minimal at best. However without a clause in our TOS to guide the mods then all we are doing is spinning our wheels

Steve,

It this a permission to include text generated by AI if the poster acknowledges it ?

IMO you can't drop the distinction between generated and mixed, although I understand that mixed must be properly defined to avoid conflicts.

 

[Tangram]

An AI-generated response (Claude, the new cool kid on the block) to a thread I started gave a fantastic answer to the thread question. But I was thankful is was flagged as AI.

 

[wil]

The more I look at this topic the more interesting it becomes...

With the advent of AI and AI detectors comes the interest in avoiding AI detection.

Here is what the Goggle AI says to my questions: how to avoid ai detection:

"To avoid AI detection in writing, focus on creating content that appears more human-like by varying sentence structure, avoiding repetitive phrases, paraphrasing text, incorporating personal anecdotes, and using a thorough editing process to add nuance and complexity, while also checking with AI detection tools to identify potential areas for improvement; essentially, "humanize" your content"

More from the robot here:

Google Search

and another site:

10 Actionable Tips To Avoid AI Detection In Writing

Avoid AI content detection in your writing with these 10 steps

surferseo.com

Hopefully the Detection AI will keep pace with the Avoidance of Detection AI.

 

[microstrip]

An AI-generated response (Claude, the new cool kid on the block) to a thread I started gave a fantastic answer to the thread question. But I was thankful is was flagged as AI.

Exactly the procedure that will kill any interesting discussion on the subject. In 10 seconds the poster pasted 20 lines of text of generic technology, not even relating it to the thread evolution. Everything is superficially addressed in three words, no one surely wants to loose its time in long post writing what can be of interest and does not apply in this encyclopaedia discussion. The kind of post who makes hard subjectivists smile ...

Apologies if you consider this a fantastic answer. It was this post that triggered my previous question to Steve.

 

[Ron Resnick]

The only things I am comfortable with so far are requiring disclosure and allowing WBF to delete posts which several AI detectors agree is AI:

The posting of artificial intelligence ("AI") generated blocks of text or information is prohibited. Notwithstanding the forgoing members may post the results of an AI search provided: 1) the search question asked of the AI is included in the member's post, and 2) the AI's answer is disclosed by the member in the post as having been generated by AI.

WBF may utilize AI detection applications to detect AI generated blocks of text or information. If WBF utilizes consistently three AI detection applications which determine unanimously that a post contains AI generated blocks of text or information then such portion of such post may be deleted.

My current thinking is that we should get some actual experience with these AI detectors and see how much controversy we have between the detectors alleging AI versus posters claiming original writing. Until we get some experience with this I am not comfortable making AI posts a basis for suspension and banning.

 

[Tangram]

Exactly the procedure that will kill any interesting discussion on the subject. In 10 seconds the poster pasted 20 lines of text of generic technology, not even relating it to the thread evolution. Everything is superficially addressed in three words, no one surely wants to loose its time in long post writing what can be of interest and does not apply in this encyclopaedia discussion. The kind of post who makes hard subjectivists smile ...

Apologies if you consider this a fantastic answer. It was this post that triggered my previous question to Steve.

Your perspective as someone who is well-versed in the topic is different from mine. I view the AI response as an augmentation to an interesting discussion. But if in some way the AI response offends people, thereby limiting their input to the discussion, that’s definitely an issue. My daughter is immersed in AI ethics at university and worked at the German Institute of Artificial Intelligence last summer, so these sorts of moral/ethical concerns get quite a bit of air time in our family.

 

[Steve williams]

I'm inclined to agree with your proposal, Steve.

I have two questions about implementation. It may be naive as I don't know much about the practical issues surrounding AI:

a) given the potential for a mixed post containing partial human and partial AI text, is your thinking that any occurence or degree of AI participation marks the message as such? I'm of the view that any occurence is problematic.

b) by what method(s) will text be determined a product of AI?

excellent question. I asked the same question question on the Xenforo forum for licensed users. At the moment there are several 3rd party Ai detectors which they are selling but in the bigger picture they realize this is an issue for everyone and there is a move afoot that they are considering a move to implement their very own AI detector plug in

 

[Lagonda]

Hello Dave,

Are you saying that bots presently are talking to each other on WBF?

(I am not talking about the BonzoBot on horns.)

Thank you.

It does explain why the BonzoBot has photographically memory !

 

[Steve williams]

gentlemen. My example as to proposed addition to the use of TOS is quite simple,. It shall not be allowed. Having said that if there is a debate going on in a thread and one member does a search and puts up a post stating , :"did a search on the topic and AI found the following for me and here are the results, I see this as OK

Ron IMO you are overthinking this as I truly believe that if it becomes a TOS the problem will quickly become minimized , if not negated.

You keep falling back on your AI detectors, These are evolving at such a rapid pace that as I said above Xenforo is contemplating adding one of their own as an add-on in which all posts are scanned. I believe the majority in the poll who voted "NO" will never use and I also believe that the small percent who voted "yes" might use it but I bet they will divulge

I happen to agree with Dave C (which you don't and are trying to convolute a situation in which there its no need.

I put up my example with the ladder system only to outline to members what our ladder system is as there is only one rule for all Terms of Service. SAnyviolation is placed on the ladder system by our mods and it takes a soft warning initially. If further recurrences then a formal warning process begins and escalates up the ladder until a temporary ban is issue. What is written in the TOS will merely state, "the use of AI on WBF is not allowed unless full divulgence by the poster at the time of the post"

We employ a ladder system that issued exactly the same in all four TOS

I still believe that if a member knowingly or unknowingly uses AI and a soft warning is issued, the behavior will stop. I believe that "we human" have the ability to process right from wrong and I believe that said member will cease and desist.

Ron you are making something simple IMO into something far too complicated. Im beginning to wonder if you re a bot.

I believe have to start somewhere and that a new TOS needs to be added and members need to be aware. I also believe that our mods act equally and without bias in all cases. Ron andd I do not moderate. That duty is done solely by the mods. Ron and I never get involved nor do we ever know what the mods are dealing with. This is purposely done becay=use we all decided that "owners do not moderate"

Finally there might be the one or few outliers who feel above the TOS and will advance up the ladder. Hopefully they will get the message or ultimately a temporary ban will occur. Surly that would stop the behavior but if not our ladder system deals with that.

I have been in touch with Xenforo and as stated on their owners forum they are already selling 3rd party AI Detectors

IMO we have chewed this to pieces and the time has come to add a TOS . We have to start somewhere and I believe my suggestion is a fair start. To answer the question about a post which suggests an AI mix I believe that informal discussions with the author should take place to discuss with him what our detectors are suggesting. I dont believe an AI mix of 2% is in any way meaningful. I do believe however that a discussion with the author is indicated for a mix with a higher percentage. Our mods are human and I believe they would know how to deal fairly with all members should such a need arise

 

[Rensselaer]

gentlemen. My example as to proposed addition to the use of TOS is quite simple,. It shall not be allowed. Having said that if there is a debate going on in a thread and one member does a search and puts up a pst stating , :"did a search lion the topic and AI found the following for me and here are the results, I see this as OK

Ron IMO you are overthinking this as I truly believe that if it becomes a TOS the problem will quickly become minimized , if not negated.

You keep falling back on your AI detectors, These are evolving at such a rapid pace that as I said above Xenforo is contemplating adding one of their own as an add-on in which all posts are scanned. I believe the majority in the poll who voted "NO" will never use and I also believe that the small percent who voted "yes" might use it but I bet they will divulge

I happen to agree with Dave C (which you don't and are trying to convolute a situation in which there its no need.

I put up my example with the ladder system only to outline to members what our ladder system is as there is only one rule for all Terms of Service. SAnyviolation is placed on the ladder system by our mods and it takes a soft warning initially. If further recurrences then a formal warning process begins and escalates up the ladder until a temporary ban is issue. What is written in the TOS will merely state, "the use of AI on WBF is not allowed unless full divulgence by the poster at the time of the post"

We employ a ladder system that issued exactly the same in all four TOS

I still believe that if a member knowingly or unknowingly uses AI and a soft warning is issued, the behavior will stop. I believe that "we human" have the ability to process right from wrong and I believe that said member will cease and desist.

Ron you are making something simple IMO into something far too complicated. Im beginning to wonder if you re a bot.

I believe have to start somewhere and that a new TOS needs to be added and members need to be aware. I also believe that our mods act equally and without bias in all cases. Ron andd I do not moderate. That duty is done solely by the mods. Ron and I never get involved nor do we ever know what the mods are dealing with. This is purposely done becay=use we all decided that "owners do not moderate"

Finally there might be the one or few outliers who feel above the TOS and will advance up the ladder. Hopefully they will get the message or ultimately a temporary ban will occur. Surly that would stop the behavior but if not our ladder system deals with that.

I have been in touch with Xenforo and as stated on their owners forum they are already selling 3rd party AI Detectors

IMO we have chewed this to pieces and the time has come to add a TOS . We have to start somewhere and I believe my suggestion is a fair start. To answer the question about a post which suggests an AI mix I believe that informal discussions with the author should take place to discuss with him what our detectors are suggesting. I dont believe an AI mix of 2% is in any way meaningful. I do believe however that a discussion with the author is indicated for a mix with a higher percentage. Our mods are human and I believe they would know how to deal fairly with all members should such a need arise

Steve, in your 9th paragraph, last (6th) sentence it reads: "This is purposely done 'becay=use' we all decided ..."

See also posting 110; by exupgh12, point 7., "2nd example", reads: "In 2002, Jan Hendrick 'SchA9n', a physicist ..."

I am wondering if AI didn't like where this was going and has put a virus on this thread to punish us? ;-)

 

[treitz3]

For a not so complete understanding of an answer provided by A.I. -

Question - How would you write, "I changed a light bulb" on your resume?

Answer - Single-handedly managed the successful deployment of a new environmental illumination system, with zero cost overruns and zero safety incidents.

(This is a joke, in case you were wondering)

Tom

 

[Steve williams]

For a not so complete understanding of an answer provided by A.I. -

Question - How would you write, "I changed a light bulb" on your resume?

Answer - Single-handedly managed the successful deployment of a new environmental illumination system, with zero cost overruns and zero safety incidents.

(This is a joke, in case you were wondering)

Tom

@treitz3

Tom for the sake of clarification and education of all members reading this, can you please post what the ladder system is, how it works, what triggers the ladder system, how many warnings, (Informal and Formal) a member gets before a ban is impose. How long is the first ban. How many more warnings before a second ban and how many before a permanent ban, Is there anything such as spam or something egregious which would require an immediate permanent ban, I feel this needs to be memorialized here as to what you and your team do as to the ladder system.

Thank you

 

[Ron Resnick]

SIMPLIFIED REVISED DRAFT

The posting of artificial intelligence ("AI") generated blocks of text is prohibited. However, members may post the results of an AI search provided the search question asked of the AI is included in the member's post, and the AI's answer is disclosed by the member in the post as having been generated by AI.

WBF utilizes the AI detection application built into Xenforo. If this application determines that a post contains AI generated blocks of text then such portion of such post may be deleted.

 

[Ron Resnick]

General Question for Discussion:

What do people think about the use of AI for checking grammar, spelling and generally smoothing one's original writing on the "back-end"? In other words not asking AI a question and then copying and pasting blocks of text of the answer as a post on WBF (if not disclosed as required this is prohibited), but writing something original, and then copying and pasting that original writing into AI to check for grammar and spelling and intelligibility?

 

[Ron Resnick]

My answer to the above question is that I am fine to prohibit this, also. The procedure of running original writing through AI will result in a narrowing of the variance of the writing styles and personalities of writing and logic and thinking styles we see here naturally.

I would not want posts on WBF over time to read like they are converging in style and tone due to AI correction and smoothing.

So I would not add to my simplified revised draft in Post #216 an exception to protect this use of AI. I would prohibit this "back-end" use of AI as well.

 

[microstrip]

SIMPLIFIED REVISED DRAFT

The posting of artificial intelligence ("AI") generated blocks of text is prohibited. However, members may post the results of an AI search provided the search question asked of the AI is included in the member's post, and the AI's answer is disclosed by the member in the post as having been generated by AI.

WBF utilizes the AI detection application built into Xenforo. If this application determines that a post contains AI generated blocks of text then such portion of such post may be deleted.

IMO this exception, as stated, will drive many posters away from WBF.

Although I would simply ban AI generated text, I have seen places where they have rules relating the percent of user text versus AI generated text - for example, allowing a maximum of 20% of AI generated text in a post.

 

[microstrip]

My answer to the above question is that I am fine to prohibit this, also. The procedure of running original writing through AI will result in a narrowing of the variance of the writing styles and personalities of writing and logic and thinking styles we see here naturally.

I would not want posts on WBF over time to read like they are converging in style and tone due to AI correction and smoothing.

So I would not propose an exception to my simplified revised draft to protect this use of AI.

We must think about what is the interest of common people running their comments through AI? IMO interested members will want to keep their onw vivid and personal style.

But considering that a relevant part of our membership is involved in business I easily see a reason to second your firm proposal of banishing any AI participation. AI assisted promotion can easily contaminate threads.

 

[Ron Resnick]

IMO this exception, as stated, will drive many posters away from WBF.

Do you not think ChatGPT can or should be used as kind of a modern version of an encyclopedia for establishing, for example, a baseline for mutually agreed technical information?

"What is a Class A amplifier?"