Do you think that AI-generated comments make communities like Product Hunt a better place?
Kyle Frost
128 replies
We've seen a lot of growth in AI comment generators lately – auto-creating comments and replies based on available info. Do you think apps like these are positive or detrimental to community-focused platforms like Product Hunt?
Fwiw, we're aware that many folks are already choosing to use these kinds of services on Product Hunt. We're actively working on ways to address it, but I'm curious how our community feels about this type of engagement.
Please vote (and explain your thoughts)!
Replies
Raj @ Vault Vision@rajnijjer
Vault Vision
I hope that is not the case!
Share
EnVsion AI
Short answer: No.
Longer answer: A community is about people. If people instead of communicating with each other within a community delegate this essential activity to machines with generic personality then the community ultimately suffers from a lack of authenticity.
I'm excited about what we're now able to do with AI but it should not be at the expense of people forgetting to be their true authentic selves online.
Life gets boring if you let machines drive all your interactions with people.
Skinive AI: Skin Scanner, health checkup
My desire is to engage with a community that consists of genuine individuals who are actively using their accounts. If I were interested in having meaningless conversations with automated bots, I would have sought out other avenues. The primary objective of this community is to exchange feedback, share genuine experiences, learn from others, and not merely to promote one's own products.
P.S. this comment was created by Chat GPT ))
Chatfuel
Yep, also had that thought. AI can help you polish your grammar and wording, but from my experience, you can almost always see the difference between AI vs. Human-generated content, especially in a small formats like comments. So, the whole concept of AI-generated comments becomes pointless.
Even ChatGPT with all its power creates generic content without much personality behind it. And it’s good for some use cases when speed matters more than the quality of created content.
But in a long distance and in real communities quality always beat quantity :)
It's like using a prepared solution instead of thinking of it by yourself. Hope there will be an auto-check of how humanized published comments.
Isn't it the fact that comments are meant to be humane (unless spamming events) to engage while AI-generated contents are more relevant for articles and blogs?
Introwise
☝️ All of those are AI-generated comments 👇 This is the only real one (trust me)
Jokes aside, this will be a problem once the quality of those comments become indistinguishable from the real ones. It's quite possible that we'll see fine-tuned tools that can write time-relevant, insightful and emotional comments in the nearest future.
And this is probably fine, as long as these comments bring real value and are not just keyword-stuffed spam. Someone spent their time and money employing AI to write a good comment that is useful to a human user. Sounds like a good thing.
We wight need to rethink the verification and reputation systems in place though. There should be something ensuring that the community dynamics benefit the human users of the platform (sorry bots, you'll have your own ProductHunt in the future, leave this one to us)
Product Hunt
@dbanisimov Totally agree about verification and reputation. We have some more changes coming soon, but in the meantime we've just added LinkedIn verification to user profiles to help add signal to who might be more trustworthy.
Communities should be genuine just as we should. Humans only
I've noticed that the responses ChatGPT will generate are either dry or lack a personal touch, even with good prompts. I've fallen into a pattern of getting it's take on a response and then either rewriting it using the general idea or just throwing it away and writing it from scratch.
Paperguide
Launching soon!
A lot of non-native English speakers who are otherwise reluctant to comment would be able to write comments now. However, it's just a tool and will result in both helpful and spam comments. If product hunt can build an algorithm/feature to filter or deprioritize spam comments, it would make PH a better place
Daily Scrum Meetings - Notion Template
I tried using ChatGPT for comments, but I found that they lacked personality and essence. We often prioritize quantity over quality. Perhaps we are expecting too much from AI, and it should be treated as a tool rather than a substitute for human interaction.
FocuSee
@laurentiu_cotet Yeah,totally agree with you. As a non-native english speaker, AI tools could make my engagements in PH communities much easier. However, they're just tools to help me express my thoughts to a wider world.
COLLABORATE
@laurentiu_cotet Asked ChatGPT for a reply to your comment. Here is what "it?" said -> "I completely understand your point of view. While AI language models like ChatGPT are capable of generating responses quickly and efficiently, they do not possess the same level of personality and emotional intelligence as human beings.
It's important to remember that AI is a tool that can assist in many tasks, including language generation, but it should not be seen as a complete substitute for human interaction. The human touch is often necessary for creating meaningful and impactful conversations.
As with any technology, it's important to use AI in the right context and not expect it to replace human interaction entirely. By recognizing its limitations and strengths, we can use AI as a tool to enhance our communication and productivity, rather than relying on it exclusively."
Hummin'
@laurentiu_cotet @dipak_sisodiya impressive. I was looking for where the ai generated comment ends and where your human-generated comment starts for a long time.
Boolvideo
@laurentiu_cotet @inc_gemoo Me too!!! I am the same as you. Besides, you've made a good product. Just go for it.
@laurentiu_cotet @inc_gemoo Absolutely! AI tools can definitely help bridge the language barrier and expand our reach. But at the end of the day, it's the quality of our thoughts and ideas that truly matter. Keep on expressing yourself!
AI-generated comments lack authenticity and sincerity, which are essential for building a meaningful community. Genuine human interaction is what sets things apart.
(Amusingly the above was made by an AI comment generator)
Hummin'
@maxwellcdavis wow I didn't notice until you mention it. which tool is it?
The biggest problem I have with it is that it breaks the connection between response length and effort.
Before AI generated comments if you wrote a block of words I at least knew you'd invested time and effort in preparing it (unless it was obvious copypasta) - and so I'd gladly invest time and effort in reading what you wrote.
Now there's no longer a connection between response length and effort - it's going to put people off reading a more thorough response - which in turn is going to put people off writing one.
The only way I can see to handle this is to let the community moderate itself - i.e. downvotes on comments.
In my ideal world PH would keep a record of that downvote / upvote ratio, and every user could configure a threshold below which they don't even see the post. i.e I might choose to only see comments from users with a high upvote ratio - automatically filtering out those that the community collectively has marked as too negative for whatever reason.
Something concerns me and i am not ready for it
I want a community with REAL human beings behind their accounts.
If I wanted to chat randomly with bots I would've done that elsewhere.
The whole idea behind this community is to get feedback, share REAL experiences, and learn from others, besides promoting your product.
@chris_sarca exactly. to me it totally defeats the purpose of having a comment section to begin with. It strips away the human experience
@roshen_earnn Agreed! 🔥
Why everyone talking about AI and forget the border line? Where human need and AI need, it's different context.
That sounds like an highway to spam.
I wanna hear people's thoughts and opinions. But I get that not everyone's confident in their English, so they might use ChatGPT to help them express themselves better. IMO that's okay, as long as they're still saying what they really mean. But if they're just letting ChatGPT do all the work and not adding their own flavor to it, then it's like, what's the point of having a community at all?
Mediocre AI comments don't lead to constructive dialogue or meaningful collaboration. Maybe it's time to look at some all-human platforms instead...
Mammoth 2
The formality and wordiness of generated comments makes them less appealing and out of place on Product Hunt.
One [obviously] AI-generated comment can erode trust and interest in all the rest of the comments on a given post, and so they definitely represent risk.
That said, AI editing tools that improve clarity, increase terseness, and help people create more muscular prose can be a net benefit.
Thus, I'm in the middle — and generally if the engagement is high quality, additive, and shared from a place of integrity, _how_ the content is produced is somewhat less important. But if AI comment generation is used simply to grow karma or artificially boost engagement, then that should be against the community guidelines.
As it is, some human commenters copy and paste (or have a script?) the same comment (or a similar variant) everyday, on every product — just creating noise. That kind of engagement is also unnecessary.
Perhaps Product Hunt could invest more in co-moderation tools so we can improve the quality and insightfulness of discourse across Product Hunt?
I know that comment badges and Comment Prompts are meant to help with that, but that's on the Maker side. Perhaps more tools (e.g. community choice selections? or "Flag as AI generated"?) beyond upvotes could help?
Product Hunt
@chrismessina you can now flag as AI using the 'report' button on launch comments "Artificially generated (e.g. ChatGPT)"
Mammoth 2
@kylefrost oh dope! Nice!
Hummin'
@chrismessina @kylefrost If a reply is reported as 'Artificially generated', will it be deleted or are there any other penalties for the user?