How do you all feel about bots?
I’ve seen a gpt powered summarization bot pop up recently. Do you find this useful? Do you hate this?
Do you think bots serve any useful purposes on this website or do you think we should ban all bots? Should we have a set of rules for how bots should interact - only when called, needing to explicitly call out they are a bot on their profile, etc?
I’d love to hear your thoughts
You must log in or register to comment.
Personally - I think any bot that could be straight Lemmy functionality shouldn’t exist but that said, I think good ground rules would be :
- Bots should be clearly prompted by a command
- Bots should not act in a community without mods from that community being contacted first
- Bots should minimize the space they take with their messages (Example: Info on how to contact its creator should be in the bot bio rather than in every message)
- Bots should say who made/hosts it
I also vote for these rules!
I agree - these seem to me like very common-sense rules.
Do spoiler foldouts maintain their functionality across UIs, either directly or in essence (eg. popup instead)?
Part of me wishes that Lemmy also had spoilers that reveal in place, but foldout spoilers have some functionality that makes me appreciate having both on hand. I’d bet bots could benefit from using that to minimize visual space if we go through with it.
These rules seem great honestly. The main bot that comes to mind is the TL;DR bot, which one could easily prompt for in a post if they want a TL;DR, if those communities want to enable it for that specific community. Eventually, a list of promptable bots could pop up in one of the instances so that people know which bots are available to be prompted. Alternatively, someone could make a website to list them or something. I can see there being a healthy bot ecosystem forming based on people’s needs.
Since we have more control over the source code, I think eventually what would be nice are community plugins to replace some of the functionality of these bots. For example, a plugin could de-AMP a link, or could provide a banner indicating the rules on a post. If someone really wanted to, they could make a plugin to auto-generate summaries of articles too and include it somewhere in the UI. Since these rules are for Beehaw specifically, I don’t think bots which create new posts are that relevant, since there aren’t really any niche-specific communities (like a bot which posts changelogs for a game or something), just broad communities.
Any bots not clearly labelled as bots should be given a warning, then banned from the instance in my opinion. The bot setting exists for a reason, bypassing it indicates that the bot author is not willing to respect the rules of the communities the bot is posting in.
If they’re informative and/or helpful, I don’t mind bots. If they’re those stupid pointless novelty bots that were plaguing Reddit, they can go away.
The grammar bots were so annoying! I love good grammar as much as anyone, but really, what help are we actually adding to the world with the they’re/their/there bots, the your/you’re bots, the payed/paid bots, etc. I really can’t imagine those changed anyone’s behavior or spelling.
I’m not completely against those, they sometimes made me edit a comment, and can be educational to both native speakers and those learning the language.
However, it’s not nice to force them upon people, it should be each user’s choice whether they want those tips or not, so I’d say: maybe, but not for Beehaw (unless maybe for some “learn-[some_language]” community).
deleted by creator
Honestly, as a programmer, I’d like the freedom to share bots that can benefit the community. Although, I do think that there should be measures in place to ensure bots don’t degrade the quality of the community.
Maybe there could be a community on Beehaw where people could post about their bots and associated commands, so we could learn how they could be called into threads where they would be helpful?
This could be added to docs.beehaw.org ;)
Please keep the bots to a minimum.
Approved bots that the admins manually review the use cases for is absolutely fine.
I just don’t want things to revert back to reddit days where I’m constantly BLOCKING new novelty bots that are absolutely freaking useless and add nothing to a conversation.
Also; PLEASE; implement the following ideas into a(n) agreement/covenant for bot operators; I quote this directly from the Tao of IRC:
The master Nap then said: “Any automata should not speak unless spoken to. Any automata shall only whisper when spoken to.”
This philosophy makes sense for IRC, but how would this work on Beehaw/Lemmy? You have to DM a bot to interact with it? How would people even know it exists? In IRC there is a list of users in the channel you can scan for helpful bots. I’m failing to see the equivalent with Lemmy.
Comment bots are mostly fine so long as they are clearly labelled, don’t take up unnecessary amounts of space, have clear purpose and add value to an article or discussion. So stuff like TLDR, Piped, Wiki bots are fine. Stuff like GROND, GPT (even though it’s cool we have a Masto feature that does that), Anakin, Musk bots aren’t useful here imo.
Post bots, I’m kind of on the side of I’d rather not see them, I like talking about articles with the user who posted it. I won’t be too upset if they end up allowed, though. A whitelist, or a strictly enforced guideline would be acceptable for me.
The TLDR bot has now been disabled as per the decision of Beehaw. Contact your favorite community mods if you’d like to change that.
Thanks, but “I’m fine with it” doesn’t necessarily mean I would miss it if it’s gone.
I say no to bots. I see the utility in some of them, but beehaw is only one instance, and I’d love to keep it an instance that I know is full of actual people. Makes me want to engage in conversation much more. Besides, if you want a bot filled experience you can engage in one of the instances that allow bots. If bots are allowed, I want them to be very clearly labeled. I want to know in one glance if I’m speaking to a bot.
Agreed. I find a lot of them to be irritating and the rest of them pretty useless and annoying.
beehaw is only one instance, and I’d love to keep it an instance that I know is full of actual people.
That’s an insightful way of putting it that didn’t come to mind.
I think part of what Beehaw uniquely offers is the drive for its own kind of instance and user culture and a closer and more organic community. Bots, save for moderator tools, admittedly detract from that kind of vibe. I could imagine that sacrificing less necessary bots, either partially or entirely, could be an important measure toward securing those aforementioned values. Federation with more Reddit-esque instances still allows us to scratch Reddit sort of itch when it comes up.
Is anyone checking the AI “summariser” bot for accuracy? I’d rather not get misleading ideas in my head from a poor summary.
Is someone checking human summarizers as well? I mean, humans make mistakes but also generally adds flavours, and can focus on things due to inherent bias. In fact, this is actually an area were bots can probably produce more factually correct and unbiased summaries than humans (depends on the quality of course).
The way past both is to actually read the article?
Erm, well, yes. That should happen too. Tends to in a good community with a range of views.
I asked a single question on a single facet of the current internet. For my own information, because I’ve found reading a range of articles about Chat GPT useful for understanding and beginning to form my own opinion on them. And rather than add any helpful information, you’ve gone down this tangent? 🤷♂️
Your “In fact” rebuttal, not needed btw, is technically true. I’m more interested in the current actual state of things with a particular bot, not a hypothetical.
Human-written posts differ in tone from the summary-bot. The bot “writes” more in the tone of an article, which tends to mean a tone of authority. That affects how the “facts” resurface in my memory. Maybe it works differently for the bright young things who’ve grown up with the internet. IDK 🤷♂️
Of course reading the articles is important. I don’t have the spoons to read every article I come across though. I know I don’t have much of a life, but still 😂 Scanning comments is a bit more like human interaction and I find that helpful in deciding whether or not to click through to the article.
And before anyone jumps in with “Then the summary bot will be really helpful to you”, please note that my question was about the accuracy of the bot and if anyone was gathering information. I will make my own observations over time but would also like to learn from others’.
The bot has now been disabled as per the decision of Beehaw. Contact your favorite community mods if you’d like to change that.
To answer your question, yes, I am checking it for accuracy as I’m the author and I’d like it to be as useful as it can be. I’d say its summary is really helpful in 90+% of cases, the rest could be better and only once I’ve seen it post a summary that wasn’t helpful at all.
Thanks for this :)
No problem! Hopefully we’ll come to some agreement with Beehaw mods/admins, you can follow the (hopefully fruitful) discussion here: https://beehaw.org/comment/858784
I think bots can have a place, but I prefer ones that have to be intentionally invoked. I’m thinking of ones like MTGCardFetcher on the Magic the Gathering subreddit, which would post links to the card on Scryfall if you formatted the card name in double brackets in your comment.
In my opinion, such bots indicate more of a need for some kind of easy “pipe” feature to integrate tools to transform a post before publishing, so that all of the tweaks can be done within the post instead of as a bot reply. For example, there could be a “MTG-ify” button that takes the text in the input box, turns the double bracketed names into CommonMark links, and then puts the modified text back into the input box.
Yeah I don’t disagree at all that it would be ideal if some of this kind of functionality could be built into the platform, but obviously that didn’t really happen at Reddit - which is why there were so many similar bots to allow subreddits to create extended functionality - and Lemmy is still new enough that contributors are still trying to fix major issues and get basic functionality working properly. In the meantime bots could fill some gaps, although I lean toward using them very sparingly.
One of the things I like about Beehaw is the lack of bot posts in every thread. Personally I think all bots should be banned because it eliminates some unwanted spam, but a good compromise for me is that bots be explicitly labeled, and can only respond to a trigger command. Nothing that auto posts.
If you think all bots should be banned, then good news! On Lemmy, bot accounts are (should be) labeled as a bot, and in your profile settings you can disable seeing posts by bots.
That’s a very good setting. Thanks! My only other concern is unlabeled bot accounts but I don’t know if that’s a rampant issue or not.
Yes, I think it’s important to have a “bots must be marked as bots” rule.
Second this, for Beehaw.
Users can already follow any community from another instance with autoposting bots. With the right interface, users can even merge posts from no-bot and yes-bot communities to create their own customized experience.
I dislike content that has been auto-posted by bots. I treat it like spam instead of genuine content.
I would love to see a “bot” flag and a parameter on your profile to not show any “bot” content.
I guess people who make bots are scared that the Lemmy platforms would eventually stop seeing activity because of a lack of content. But I think that if there were little to no activity, perhaps people would be posting more. I doubt that flooding the platform with auto-generated content or auto-forwarded content actually helps with encouraging people to stay.
IIRC there is already a bot flag on profiles, though it relies on bot-makers manually setting it and as far as I’m aware you can’t automatically block all bot users (though I haven’t tried every single Lemmy app).
I don’t want bots on Beehaw. Either unknown ChatGPT generated comments or bots that just listen to keywords and hey heres a Wikipedia link type. I want discussion from real, good, people with opinions. Not a bot with useless commentary I could just Google(Kagi) instead. Rules around this type of bot is okay, this isnt gets into rules lawyering and favoritism. My vote is no to bots.
Honestly the only bot I’ve actually found myself missing is the metric/imperial conversion one, makes talking with Americans a lot easier!
98% of bots are crap. The problem is that people have different opinions as to which 98% of them is the crap portion.
Absolutely any bot needs to self-identify.
I’m of the opinion that bots are okay if:
-
They provide value to the community - A news-bot seems to be well received at tucson.social and it helps people get all their Tucson updates in one place without having to share it themselves.
-
They assist with moderation. Auto responding to new posts that reminds thread participants of the rules could be one use-case.
-
They enhance the dialogue of the thread or provide useful and important corrections. Perhaps there’s a bot that looks up species names and provides useful links in a reply of a zoological based post? I say that’s great and what we want!
As for ChatGPT bots:
- All bots must disclose they are a bot.
- All bots must not fake engagement. As in, it’s okay to be other bots because of their relatively strict use-cases and minimal ability to hallucinate and no ability to respond to further queries. ChatGPT makes it appear as if it’s a person at times and can be subtly wrong - we have people that do that just fine.
- ChatGPT content should go into their own relevant subs. A MachineLearning community might be good at first, but perhaps eventually a dedicated LLM/ChatGPT Writes type community would eventually be needed for peoples more creative impulses. It’s not exactly relevant for someplace like tucson.social, but might be for a place like BeeHaw.
Auto responding to new posts that reminds thread participants of the rules could be one use-case.
IMHO those pinned top messages in Reddit were a stopgap for dealing with highly diverse communities and moderation styles “on a single instance”.
Again in my opinion, the Fediverse would benefit more from having consistent rules per-instance, with only sub-rules on a community level. Both of these should be made easily discoverable to all participants of a “community@instance” directly through their interface (web or app), making the pinned top messages unnecessary.
Communities with “highly diverse moderation styles”, should rather stay on separate instances with similar moderation styles, making it easier for mods to apply a consistent ruleset, for users to decide which instances to follow, and admins whether to federate or not. There already exist interfaces (both web and app) to merge communities from multiple instances if the user so wishes to (at their own risk, but again IMHO the rule differences should be handled by the user’s interface).
Ideally, I think that users should be able to use an interface of their own choice to merge comments on a matching post from multiple instances or groups of instances (federated), interacting in whatever style they choose without interfering with users who didn’t choose that style.
Particularly in the case of Beehaw, which has a consistent set of “rules but not rules” for all communities, I think those messages would only add clutter.
This makes sense, but I think that Lemmy just has this same problem on a different scale (between instances rather than between communities). The problem we have seen sometimes is folks seeing Beehaw posts in the All feed of their home instance and coming in and commenting/posting without knowing what/who we are and without engaging with the sidebar or any of our docs. And some federated sites make it difficult to even tell that you are seeing a post from another instance (I’m looking at you, Kbin). The vast majority of the time it isn’t a huge problem, but it does mean that the mods are having the same conversation over and over because some folks aren’t aware of the vibe of the place where they’re posting.
Now obviously an automatic bot comment would be a band-aid, and I suspect not a particularly effective one (Lemmy doesn’t provide the ability to sticky comments). It would be ideal if there was some functionality built into Lemmy itself to remind users of the instance they are about to post in, and the rules of that instance.
an automatic bot comment would be a band-aid
I’d kind of rather they get this as a direct message or even better a warning on the UI they interact with our site on (we cannot enforce the latter).
Yeah, I had been thinking that a better solution would be some built in Lemmy functionality - I think Reddit actually had a reminder like this on posts, but not comments - but between apps, third-party front ends, and federation with other services like Mastodon and Kbin it probably wouldn’t reach the folks that would most need to be reminded of the rules of the place they’re posting in.
-
I’m in favor of the guidelines listed by @Lionir for bots operating on Beehaw. Particularly the part about contacting community mods before deploying - it feels like the nice thing to do before adding new wrinkles to the moderation workload (which includes monitoring discussions about the appearance of the bots). That also provides an opportunity for a discussion within the community to engage with, or pre-emptively disengage from the bot account should they choose, rather than having to do it in the spur of the moment.
Direct link to Lionir’s comment on bots - it’s very near your comment in the discussion right now, but adding a link just in case these drift apart.
I appreciate you doing that, thank you
