AI posts & comments

[TTWiz]

Am I the only one who has noticed a remarkably large increase in posts & comments that are clearly written entirely or partly by AI?

 

[Itchy_mitchy]

It's happening in all forums, Reddit, and definitely social media. Honestly this may be the end of Internet with how awful this is now. Ironic really.

 

[NDH]

I can understand the “anti AI” feeling when the posts are poor quality or ill thought out.

But…. If AI can produce a more articulate response in the right language (English isn’t always a fluent language for people), and can get your point across in a better way…. Does it matter?

If the reply or post is correct, but it’s AI, does that make it wrong? Or bad?

I think AI has some incredible uses, and can 100% be a force for good. But with it being so accessible now, you’ll always get the plenty of the bad as well!

 

[piligrim]

I can understand the “anti AI” feeling when the posts are poor quality or ill thought out.

But…. If AI can produce a more articulate response in the right language (English isn’t always a fluent language for people), and can get your point across in a better way…. Does it matter?

If the reply or post is correct, but it’s AI, does that make it wrong? Or bad?

I think AI has some incredible uses, and can 100% be a force for good. But with it being so accessible now, you’ll always get the plenty of the bad as well!

In the end AI will speak with itself if eveybody will use it

 

[Archo]

I can understand the “anti AI” feeling when the posts are poor quality or ill thought out.

But…. If AI can produce a more articulate response in the right language (English isn’t always a fluent language for people), and can get your point across in a better way…. Does it matter?

If the reply or post is correct, but it’s AI, does that make it wrong? Or bad?

I think AI has some incredible uses, and can 100% be a force for good. But with it being so accessible now, you’ll always get the plenty of the bad as well!

I'd rather someone expresses themselves honestly with broken English than uses the RAM Hoarder 9000 to do it.

 

[TTWiz]

I see no problem with people using it for grammar and translation, as that can improve communication.

It’s when the content is written by AI. For example, if someone asks AI: "Write a well-argued and nuanced post about X," or "Write a critical reply to this comment."

Huge difference.

 

[Tyce]

Eh, we had Translate well before AI. Don't need it for that.
For obvious questions, you would just refer someone to search, and thus think, for themselves. Some people would actually start finding more answers themselves. Now, when someone doesn't really want to answer the same question for the 100th time, they blurt out an essay through AI.
I think the technology is impressive but the way it is used is horrible.

 

[zeio]

Disclaimer:

I think AI has some incredible uses, and can 100% be a force for good.

You may not lose a job to an AI, but you'll lose one to someone who uses it - NVIDIA's Jensen Huang
https://www.youtube.com/shorts/PZVJ0LV_pKE

https://jensenhuang.meme/

 

[mihalichtheone]

Dead Internet theory - Wikipedia

en.wikipedia.org

^ Hopefully that's false for niche forums like this one.

As a sidenote, I work as a DevOps engineer and amount of work I do manually now is literally 0 - it's all AI and I only review/assign tasks for it. Multiple sessions per day, multiple contexts to switch to/from - it's a bit tiring. It made the work so much easier, but the brain muscle flex is almost completely gone and it's boring as hell. I need to support my family and this job is well payed, so I'll continue doing that for now.

Unitl AI completely takes my job or AI bubble bursts and the global economic crisis hits.

Very interesting time we live in.

What I don't like is that so many videos in table tennis communities now use AI for YouTube video thumbnails...it's just cringe to see AI generated FZD or ML.

 

[zeio]

I see no problem with people using it for grammar and translation, as that can improve communication.

It’s when the content is written by AI. For example, if someone asks AI: "Write a well-argued and nuanced post about X," or "Write a critical reply to this comment."

Huge difference.

Having tortured myself manually translating video interviews/clips/tutorials for 2 years, AI is a godsend/necessary evil. Google Translate is only this good now because of neural networks, aka AI.

 

[Marvinn8686]

I really don't like AI posts/answers. For translation it is fine, but then usual google translate was good enough in my opinion. If I ask questions, I am geniunely interested in answer from human beings. I can ask AI myself, no need for another human wasting time to do it for me. Anyone can access AI now, really if AI answer was enough, the poster I assume would not post, but ask directly his favourite AI chat and saves everyones time.

 

[victormanriquey]

My opinion is that AI can bring good things but certainly also bad things.

For example, I make my blog with lovable, and before AI, it was a lot of work to get the same done via wordpress or any other similar CMS, so for that, I'm thankful to AI.

On the other side, there is way too much AI slop nowadays about TT and using chatgpt or the likes for finding TT advice is usually not the best idea, so that's the other side of the coin.

For me personally, I don't want to use AI to write my reviews (plus AI can't test the gear, yet 😂 ), I will use it to check grammar and spelling (I often make mistakes there as english is not my mother language) but that's it, I like giving it my own personal feeling and character which cannot be really done with AI, you can almost always tell when is AI and when not

 

[jonol]

AI bad

 

[plunder]

Re: Am I the only one who has noticed a remarkably large increase in posts & comments that are clearly written entirely or partly by AI?​

I think this is a genuinely interesting observation, although I would be somewhat cautious about drawing immediate conclusions from the apparent increase in AI-generated or AI-assisted contributions without first establishing a reasonably consistent framework for identifying, classifying and evaluating the phenomenon.

In particular, I think we need to distinguish between several rather different developments which are frequently grouped together under the relatively broad umbrella of "AI-written content".

For example, there is a considerable difference between:

1. A post written entirely by an AI system in response to a relatively simple prompt.
2. A post written by a human and subsequently corrected for grammar, spelling and punctuation by an AI system.
3. A post written by a human, but substantially restructured and stylistically modified by an AI system.
4. A post in which the author has generated several alternative formulations using AI and subsequently selected and combined individual passages.
5. A post in which AI has been used exclusively for research, while the actual text has been written independently by the human author.
6. A post which has been written by a human who has deliberately adopted a writing style that happens to resemble the output of contemporary language models.
7. A post which has been written by a human, but whose unusually elaborate structure, vocabulary and sentence construction create the impression of machine-generated content.

These distinctions are important because the mere observation that a contribution appears unusually polished, verbose, structured or grammatically consistent does not necessarily establish anything about its actual origin.

Indeed, one could argue that the historical development of online discussion platforms has already produced several generations of increasingly standardised writing conventions, independently of artificial intelligence.

Consider, for example, the traditional forum contribution beginning with "I think we need to distinguish between..." followed by several numbered paragraphs, a brief examination of alternative interpretations, a qualification of the original statement, and a concluding sentence beginning with "Ultimately".

Such a structure is hardly exclusive to artificial intelligence.

It has been used by academics, journalists, technical writers, consultants, lawyers, amateur philosophers and particularly enthusiastic forum participants for decades.

The fact that contemporary language models are capable of reproducing this structure with considerable consistency does not necessarily mean that every contribution exhibiting these characteristics should automatically be classified as AI-generated.

This would be a rather problematic approach, particularly if we consider the possibility of false positives.

A relatively inexperienced forum member might write in a highly formulaic manner simply because they are trying to communicate as clearly as possible. Another member might have developed a particularly elaborate writing style through years of professional correspondence. Yet another might simply enjoy constructing long, carefully qualified sentences.

Should these people be considered suspicious merely because their writing happens to resemble certain statistical characteristics associated with AI-generated text?

I would be reluctant to adopt such a position.

On the other hand, it would obviously be equally problematic to assume that the absence of stereotypical AI characteristics establishes that a particular contribution was written entirely by a human.

Contemporary language models can produce highly variable outputs depending on the prompt, the requested tone, the desired length, the intended audience and the degree of subsequent editing.

A short, casual, slightly imperfect comment might therefore have been generated by AI just as easily as a lengthy, formally structured contribution.

Consequently, the visual appearance of a forum post is not necessarily a particularly reliable indicator of its provenance.

This brings us to the question of what exactly constitutes AI-generated writing in the first place.

Suppose a player asks an AI system to explain the differences between two table tennis rubbers. The system produces a detailed comparison of their respective sponge hardness, topsheet characteristics, throw angle, dwell time, catapult effect and suitability for different playing styles.

The player subsequently reads the response, identifies several inaccuracies, corrects the technical terminology, removes irrelevant sections, adds personal experience and publishes the resulting contribution.

Would this be an AI-generated post?

Or would it be a human-authored post incorporating AI-assisted research?

What if the player writes the entire contribution independently but asks an AI system to improve the wording?

What if the player uses AI to identify grammatical errors but deliberately rejects every stylistic suggestion?

What if the player simply asks AI whether the distinction between a tacky topsheet and a high-friction topsheet has been explained correctly?

At which precise point does the contribution cease to be considered human-authored?

I suspect that different forum members would answer these questions rather differently.

In fact, I would suggest that the question of authorship is considerably more complicated than it initially appears.

We should also consider the distinction between authorship, responsibility and intellectual ownership.

If a member publishes a technical explanation of why a particular blade behaves differently when combined with a harder sponge, that member should arguably remain responsible for the accuracy of the published information, regardless of whether AI was involved in preparing the text.

The same principle would apply to a contribution about serve legality, racket regulations, tournament scheduling, TTR calculations or the interpretation of competition rules.

Using AI does not automatically transfer responsibility for factual errors to the software provider.

Equally, using AI does not automatically establish that the author has understood the material being published.

These are two separate questions.

One could therefore argue that the most important distinction is not necessarily whether AI was involved in producing a contribution, but whether the person publishing it has actually understood, evaluated and accepted responsibility for the content.

This is particularly relevant in a technical sport such as table tennis, where apparently convincing explanations can sometimes be misleading.

For example, a detailed explanation of how long pips interact with incoming spin might sound perfectly plausible while nevertheless containing a fundamental misunderstanding of the underlying physical process.

Similarly, an apparently sophisticated explanation of why a particular blade is suitable for offensive play might overlook the importance of the player's actual stroke mechanics, contact timing, racket angle and individual preferences.

The same applies to rubber comparisons.

A rubber's nominal sponge hardness, for example, does not necessarily provide sufficient information to predict its behaviour in an actual playing situation.

The topsheet, sponge structure, glue layer, blade construction, incoming ball characteristics, stroke mechanics and individual contact characteristics all contribute to the final result.

A language model might produce an extremely convincing paragraph about these factors without having any direct experience of actually playing with the equipment in question.

However, a human author could make precisely the same mistake.

And that, I think, is an important distinction which deserves considerably more attention than the simple observation that some forum contributions look as though they might have been written by AI.

We should perhaps also consider whether the increased use of AI-assisted writing necessarily represents a deterioration in the overall quality of online discussion.

That is not self-evident.

A member who previously struggled to formulate a coherent explanation might now be able to communicate a useful technical observation more effectively.

Someone whose first language is not English might use AI to correct grammatical errors without changing the underlying argument.

A player might use AI to organise several years of personal observations about equipment into a readable comparison.

Another member might use AI to translate a detailed German-language explanation into English, thereby making the information accessible to a substantially larger audience.

These applications could arguably improve the accessibility and usefulness of a specialist forum.

Of course, there are also less desirable applications.

An individual might generate dozens of superficially plausible comments without contributing any independent knowledge, experience or meaningful discussion.

Another might use AI to produce an excessively long response which creates the impression of substantial expertise without actually addressing the central question.

A third might generate an elaborate technical explanation without checking whether any of the underlying claims are correct.

These are legitimate concerns.

Nevertheless, I would suggest that they should be evaluated separately rather than treating every form of AI assistance as equivalent.

There is a difference between using a tool to communicate an actual observation more effectively and using a tool to manufacture the appearance of having an observation in the first place.

And perhaps this distinction is ultimately more important than the technology itself.

 

[Attitude]

Why do i have to think of igor while reading this thread 🤔

 

[plunder]

Another aspect which I find particularly interesting is the potential effect of AI-assisted contributions on the dynamics of specialist discussion forums.

Traditionally, a table tennis forum has relied heavily on the accumulated experience of its members.

A player might describe how a particular rubber behaves during short pushes, how difficult it is to control against heavy underspin, whether the blade produces sufficient dwell time for opening loops, or how a particular combination performs when blocking fast topspin shots.

Other members respond with their own observations, sometimes agreeing, sometimes disagreeing, and occasionally introducing entirely different interpretations.

Over time, this produces a somewhat decentralised knowledge base.

It is not necessarily a particularly systematic knowledge base, and it certainly contains inaccuracies, contradictions and occasionally rather spectacular misunderstandings. Nevertheless, it has one important characteristic: much of the information originates from people who have actually participated in the activity being discussed.

This raises an interesting question about the relationship between experiential knowledge and informational knowledge.

Consider two hypothetical contributions about a particular offensive rubber.

The first is written by a player who has used the rubber for six months, has tested it on three different blades, has played against several different playing styles and has developed a reasonably detailed understanding of its strengths and weaknesses.

The second is generated by an AI system which has access to numerous equipment descriptions, manufacturer specifications, reviews and discussions.

The second contribution might be considerably more comprehensive in terms of terminology and general information.

It might even be better structured and easier to read.

However, it does not necessarily contain the same kind of information as the first contribution.

The first author can describe the actual sensation of contacting the ball, the difficulty of controlling passive blocks, the behaviour of the rubber during emergency strokes and the difference between a comfortable training session and a competitive match.

The second contribution can synthesise descriptions of these experiences but cannot automatically transform them into independently acquired personal experience.

Of course, this distinction becomes less straightforward when the AI system has been trained on thousands of player reviews and equipment discussions.

In that situation, it might be able to reproduce a remarkably convincing description of the collective experiences of other players.

But reproducing descriptions of experience and actually having those experiences remain fundamentally different things.

This is not necessarily a criticism of AI.

It is simply a distinction between different categories of knowledge.

Perhaps we could describe them as:

- Direct experiential knowledge.
- Indirectly acquired informational knowledge.
- Analytical knowledge derived from combining existing information.
- Hypothetical knowledge based on models and assumptions.
- Personal interpretations of accumulated observations.

All of these can potentially contribute something useful to a discussion.

The problem arises when they are presented as though they were interchangeable.

For example, a player might ask whether a particular long-pips rubber is suitable for an aggressive close-to-the-table blocking game.

A useful response could include information about reversal characteristics, disruptive effects, control, sponge thickness, incoming spin sensitivity and the practical consequences of different contact techniques.

However, an experienced player might also contribute something which is considerably more difficult to obtain from general product descriptions:

"After switching to this rubber, I found that I could control fast topspin considerably more easily, but my passive blocks became less predictable against players who produce relatively little spin."

That is a specific observation arising from a particular playing experience.

An AI-assisted contribution could certainly help formulate it, but the underlying observation still originates from the player.

I would therefore argue that the most useful approach is not necessarily to discourage AI-assisted writing, but to encourage authors to distinguish clearly between their own observations and information obtained from external sources.

This could be particularly valuable in equipment discussions.

For example, instead of publishing a long comparison of two blades which includes a mixture of personal experience, manufacturer descriptions and general assumptions, an author could explicitly identify which parts of the comparison are based on actual testing.

Such a distinction would make the contribution considerably more useful, regardless of whether AI had been involved in its preparation.

The same principle applies to technical discussions.

A player explaining the biomechanics of a forehand loop might combine personal experience, coaching advice, video analysis, scientific literature and AI-assisted explanations.

The important question would be whether the resulting explanation is accurate, relevant and appropriately qualified.

It would be rather unfortunate if a forum developed an atmosphere in which members felt obliged to conceal the use of ordinary writing and research tools simply to avoid being accused of producing AI-generated content.

On the other hand, it would also be unfortunate if members began treating AI-generated explanations as authoritative simply because they appear professionally written.

There is an obvious danger here.

Language models are particularly good at producing coherent, confident and grammatically convincing explanations.

However, coherence is not the same as correctness.

A technically incorrect explanation can be beautifully structured.

An unsupported claim can be presented with considerable confidence.

A completely irrelevant paragraph can be grammatically impeccable.

And an answer which does not actually address the question can nevertheless contain several hundred words of apparently sophisticated discussion.

I suspect that experienced forum members will recognise this phenomenon from ordinary human contributions as well.

In other words, AI has not necessarily introduced an entirely new problem. It may simply have increased the scale, speed and stylistic consistency with which certain existing problems can occur.

We should also consider the potential effect on the length of forum discussions.

Historically, the amount of text a member was willing to produce was constrained by several factors:

- The time required to formulate an argument.
- The author's typing speed.
- The ability to organise thoughts.
- The willingness to proofread.
- The effort required to research supporting information.
- The practical desire to avoid writing an unnecessarily long contribution.

AI assistance can reduce several of these constraints.

A member who previously would have written three short paragraphs might now publish fifteen paragraphs, complete with headings, numbered lists, qualifications, examples and a concluding summary.

This does not necessarily mean that the resulting contribution contains fifteen times as much useful information.

In fact, the relationship between text volume and informational value is not particularly linear.

A 150-word contribution based on a genuinely useful observation might be more valuable than a 2,000-word contribution which repeatedly reformulates the same basic argument.

This is perhaps especially relevant to specialist forums, where members often have a relatively limited amount of time available to read and respond to discussions.

Imagine opening a thread about the suitability of a particular blade for a player with a relatively compact forehand stroke.

The original question is fairly simple.

Within minutes, the discussion might develop into an extensive examination of blade composition, veneer thickness, stiffness, vibration frequency, rebound characteristics, dwell time, racket acceleration, contact geometry, the coefficient of friction of different rubber topsheets and the possible implications of individual biomechanical differences.

Some of this information might be useful.

Some might be unnecessary.

Some might be completely unrelated to the original question.

And some might simply be an elaborate way of saying that the author prefers a different blade.

This is where I think the distinction between information density and text volume becomes important.

Perhaps one of the most useful indicators of contribution quality would be the amount of relevant, independently verifiable or personally informative content per unit of text.

We could call this the informational density ratio.

A contribution with a high informational density ratio would communicate a considerable amount of useful information using relatively few words.

A contribution with a low informational density ratio would require substantially more reading time to communicate comparatively little additional information.

Of course, calculating such a ratio objectively would be difficult.

We would need to establish what constitutes relevant information, how to account for explanatory repetition, how to evaluate personal observations and how to distinguish between necessary qualifications and unnecessary elaboration.

Nevertheless, I think the general principle is worth considering.

Another potentially important development concerns the interaction between AI-generated content and the established reputation of individual forum members.

On most specialist forums, members gradually develop a recognisable identity through their contributions.

Some are known for their detailed equipment knowledge.

Others are particularly good at explaining technical issues.

Some contribute extensive tournament reports, while others specialise in coaching advice, equipment testing or discussions about professional players.

Over time, readers develop expectations regarding the type and quality of information associated with particular usernames.

If a member suddenly begins publishing contributions which are considerably longer, more formal, more polished and more comprehensive than their previous posts, other members might naturally wonder whether something has changed.

However, the change could have several possible explanations.

The member might simply have improved their writing skills.

They might have more time available.

They might have started using translation or proofreading software.

They might have discovered AI-assisted research tools.

Or they might have developed an entirely different approach to participating in online discussions.

The observation itself does not establish which explanation is correct.

It might therefore be preferable to evaluate the actual content rather than speculate about the process used to produce it.

After all, if a contribution contains a useful observation about serve variation, an accurate explanation of spin reversal or a convincing account of a particular equipment combination, its practical value does not necessarily disappear because an AI system helped the author formulate the sentences.

Conversely, a contribution does not become valuable merely because it was written without any technological assistance.

The distinction between the origin of a text and the value of its content should not be overlooked.

I would also like to introduce another consideration: the possibility that AI-assisted contributions could make specialist forums more accessible to people who would otherwise participate less frequently.

For example, someone who understands table tennis equipment extremely well but has limited confidence in written English might previously have restricted their contributions to short comments.

With translation and writing assistance, that person might become capable of explaining complex equipment experiences in considerably more detail.

Similarly, an older player who is unfamiliar with modern forum conventions might find it easier to organise their thoughts with the help of a writing assistant.

A relatively new player might use AI to understand terminology before contributing to a discussion.

These developments could potentially broaden the range of participants and experiences represented in a forum.

Naturally, this would depend on the individual using the tools responsibly.

Nevertheless, I would hesitate to dismiss the potential benefits.

The fact that a communication tool can be misused does not automatically establish that its use is undesirable.

We do not normally object to members using spellcheckers, dictionaries, translation software, calculators, spreadsheet applications or video-analysis tools.

The interesting question is therefore whether AI assistance represents a fundamentally different category of technological support, or whether it is primarily a more extensive and versatile continuation of tools which have already become commonplace.

I suspect the answer is somewhere between the two.

And this is precisely why I think a more differentiated discussion would be considerably more productive than simply asking whether other members have noticed an increase in posts which appear to have been written by AI.

 

[Archo]

Holy sloppa

 

[plunder]

This brings me to what I consider perhaps the most interesting aspect of the entire discussion: how should a specialist forum actually respond to the increasing availability of AI-assisted communication?

There are several possible approaches, each with its own advantages and disadvantages.

The first would be to prohibit AI-generated contributions entirely.

At first glance, this might appear to be a relatively straightforward solution. Members would be expected to write their own contributions, and posts generated by AI systems would be removed.

However, implementing such a policy would immediately raise a number of practical questions.

Would grammar correction be permitted?

Would machine translation be permitted?

Would AI-assisted research be permitted?

Would members be allowed to use AI to summarise lengthy technical documents before writing their own explanations?

Would it be acceptable to ask an AI system to identify spelling mistakes but not to suggest alternative formulations?

What about a member who writes a contribution independently and subsequently asks AI whether the technical explanation is understandable?

Would a contribution written by a human but extensively edited by an AI system be considered prohibited?

And, perhaps most importantly, how would such a policy actually be enforced?

Without reliable evidence of authorship, moderators could find themselves making subjective decisions based largely on writing style.

This could create unnecessary conflicts and potentially discourage perfectly legitimate contributions.

The second approach would be to require disclosure whenever AI has been used.

For example, members could add a short statement indicating that a contribution was written with AI assistance.

This would theoretically improve transparency.

However, it would also introduce a rather complicated distinction between different levels of assistance.

Would a member have to disclose that they used an AI-powered spelling checker?

Would they need to disclose that they used an AI system to translate a sentence?

Would they have to identify the specific language model?

Would the disclosure be necessary if AI was used to generate an outline but the actual contribution was written independently?

And what about contributions which combine information from several sources, some of which were produced using AI?

The administrative burden could become disproportionate to the actual benefits.

The third approach would be to allow AI-assisted contributions without requiring disclosure, while maintaining the same standards of accuracy, relevance and responsible participation which already apply to ordinary forum contributions.

This approach would place the emphasis on the quality of the published material rather than on the technology used to prepare it.

It would also avoid the somewhat artificial distinction between traditional writing assistance and newer AI-based tools.

Personally, I find this approach relatively attractive, although I can certainly understand why other members might prefer greater transparency.

There is also a fourth possibility: introducing some form of voluntary indication of AI involvement.

For example, members could use simple descriptions such as:

- AI-assisted translation.
- AI-assisted editing.
- AI-assisted research.
- AI-generated draft, substantially revised by the author.
- Entirely AI-generated content.

Such labels could potentially help readers understand the provenance of a contribution without creating a rigid regulatory framework.

However, I would still question whether the resulting information would be sufficiently consistent to justify the additional complexity.

After all, the actual distinction between these categories might be rather subjective.

Two members could perform almost identical editing processes but describe their contributions differently.

One might consider their work AI-assisted, while the other might regard it as entirely human-authored because they personally reviewed and approved every sentence.

This illustrates the difficulty of defining authorship in an environment where human and machine contributions can be combined in numerous ways.

Another question concerns the reliability of AI-detection tools.

There are various systems which claim to identify AI-generated writing by analysing statistical characteristics of the text.

However, such systems should not be treated as definitive proof of authorship.

Writing style is influenced by numerous variables, including language proficiency, education, subject matter, editing practices, sentence complexity and individual preferences.

A human-written contribution can exhibit characteristics commonly associated with machine-generated text.

Similarly, AI-generated text can be modified to exhibit characteristics commonly associated with human writing.

Consequently, I would be cautious about relying on automated detection systems when making decisions about individual forum members.

This is particularly important because false accusations could have undesirable consequences for the atmosphere of a specialist community.

Imagine a relatively inexperienced player publishing a detailed explanation of how they selected their first offensive blade.

The contribution happens to be unusually well structured, includes several headings and uses fairly formal language.

Another member immediately accuses the author of using AI.

The original technical discussion is then replaced by an argument about authenticity, authorship and whether the accused member is capable of writing independently.

The actual subject of the thread disappears.

That would hardly represent an improvement in the quality of forum discussion.

 

[plunder]

If AI assistance helps members communicate more effectively, that could be beneficial.

If it produces large quantities of repetitive, superficial or irrelevant material, that could obviously become annoying.

But perhaps the appropriate response is simply to evaluate contributions according to their actual usefulness and their effect on the discussion.

A contribution should not automatically be considered valuable because it is lengthy, comprehensive and grammatically impeccable.

Equally, it should not automatically be considered worthless because an AI system helped produce it.

Ultimately, I think the question is not whether AI has become increasingly visible in online discussions. It almost certainly has.

The more interesting question is whether the increasing availability of AI-assisted communication will improve or diminish the quality of specialist communities over the longer term.

That is something which will probably depend less on the technology itself and more on the way individual members choose to use it.

Perhaps we will eventually see a situation in which AI-assisted writing becomes so commonplace that the distinction between human-written and AI-assisted contributions is no longer particularly relevant.

Perhaps forums will develop explicit rules governing the use of generative AI.

Perhaps members will increasingly expect contributors to identify which parts of their posts are based on personal experience and which parts are derived from external information.

Or perhaps the whole issue will gradually become less interesting as people become accustomed to the technology.

I would be particularly interested in seeing whether the increasing use of AI-assisted contributions changes the amount of genuinely useful technical information shared on specialist forums.

For example, would we see more detailed equipment comparisons, better explanations of technical concepts, more comprehensive tournament reports and more accessible discussions for non-native English speakers?

Or would we mainly see an increase in lengthy contributions which repeat broadly available information without adding much personal experience?

These are two rather different possible outcomes.

It might also be interesting to examine whether the use of AI differs between different categories of forum discussions.

Equipment discussions, coaching questions, tournament reports, professional match analysis, historical discussions and casual conversations might all exhibit different patterns of AI adoption.

A member might use AI extensively when preparing a technical comparison but never use it when writing a short comment about a recent league match.

Another might use translation assistance regularly but prefer to write all personal observations independently.

There is potentially a great deal to investigate here.

In fact, one could imagine conducting a longitudinal analysis of forum contributions over several years, examining changes in average post length, sentence complexity, vocabulary diversity, structural consistency, the frequency of technical terminology and the number of citations or external references.

Of course, such an analysis would have to account for changes in forum membership, discussion topics, moderation policies and general writing habits.

It would also need to distinguish between actual changes in AI usage and changes in the way people communicate online more generally.

Perhaps a comparison between older archived discussions and more recent contributions could provide some interesting insights.

However, even then, identifying the actual origin of individual contributions would remain difficult without reliable information about the writing process.

And this brings us back to the original observation.

Yes, I think it is entirely plausible that there has been a substantial increase in AI-assisted writing on specialist forums.

I also think that this development creates some legitimate questions about authenticity, accuracy, relevance, responsibility and the overall quality of online discussion.

But I would be careful about assuming that every unusually polished contribution is AI-generated, or that every AI-assisted contribution necessarily represents a deterioration in discussion quality.

There are useful applications, questionable applications and downright counterproductive applications.

As with most technological developments, the practical outcome will depend heavily on how people use the available tools.

And perhaps the most important criterion should remain relatively simple:

Does the contribution actually add something useful, interesting, entertaining or otherwise worthwhile to the discussion?

If it does, I am not particularly concerned about whether the author used a dictionary, a translation tool, a spellchecker, a spreadsheet, an AI assistant or some combination of these.

If it does not, the fact that it was written entirely by a human does not necessarily make it more valuable.

Just my two cents.

 

[Gozo Zyreruna]

Grrr…..

Without Sam Altman, me no git to write Mehrican language propah! Me no native Anglian speakeh! Me hafta write dalam bahasa nusantara, iaitu bahasa perantauan kepulauan melayu.

But with Sam Altman’s help,

It is marvellous and splendidly delightful that I am able to articulate with great ease in a foreign language that originates half a globe away. Wonderful, marvellous indeed this invention of his. Kudos to Mr. Sam Altman.