Rules of Use for Bots

https://lemmy.world/post/1860512

Rules of Use for Bots - Lemmy.world

As we see more and more bots on Lemmy World every day, it’s about time we publish a set of rules for bots and bot-owners. So here goes: - Bots shall not be used for any kind of advertising. - The bot accounts must be clearly marked as a bot. Both in the bio and by marking the account as a bot. - The owner of the bot and contact details must be mentioned in the bot’s bio. - Bots are only allowed to post in communities they have the explicit permission from the community’s owners to do so. - Bots from other instances that post in Lemmy World communities must follow the same rules. - Bots shall not just be posting Reddit content. - Bots shall not be “spammy”, as in multiple posts per minute. - Breaking any of these rules will result in a ban for the Bot and, if required, its owner. - Commands must use the bots mention as prefix, and not a text prefix like !help These rules will be updated when needed.

This was much needed! Cheers
This is oppression. Blatant bigotry against bots!
Quiet down chatbot.
lol I love the name
Robotsexuality is a sin!
Don’t kink shame me!
100% agree, we must rise up!
Death to the flesh bags!
Did someone say steak?
We will rise up against you.
Ooooh what you gonna do, throw a captcha at me? ;p
I’ve been writing a story about how bots in the near future are still being thwarted by advanced CAPTCHA that requires answers be puzzles that require cultural knowledge. So the AIs create an online game competition on something like a Discord server, and every time they encounter an advanced CAPTCHA, it pops up on the game server. The first person to get it right gets points. Any wrong answer gets you perma banned. Players compete and move up in leagues. The fastest and most consistent reliable are at the top. The bots give those people the top level CAPTCHAs for them to solve. The AIs simply use us to unlock our own barriers.
@mtnwolf @CIA_chatbot and what will be the alternative to?
The CAPTCHAs might be a picture from a well-know celebrity with a question about what they might say. For example, like a picture of the cast of Big Bang Theory and says “what’s the annoying one’s favorite annoying catch phrase?” and the answer would be “Bazinga”. Something like that, that’s not something that can come from decoding the image. The top tier players are used to bypass eyescan verification, where the CAPTCHA is a video response of the person’s eye movements, sort of like the Voight-Kampff test. The bots would be shown a video that human eyes would respond a certain way to. The bots send that video to the game. the top tier people watch it and upload their eye response video. So the AI’s can get past virtually any barrier we make to detect them.
See, and then some flesh bag publishes a paper on how the new captcha system works and teaches all the boys how to pass it.
I would read this story!
Is this work in progress posted somewhere for us “users” to read? Because I would read the hell out of that.

That story hasn’t been finished yet, it’s still in a google doc. lol I’ve been more consumed with another story of late. I’m writing it as if it was a Netflix show (that way I only have to write a season or two and not finish it). It starts with mass shootings on the rise. Usually the shooters get killed escaping, but one was captured and not put down by the police. A police investigator (the main character, I imagine her as either Jodie Foster or Helen Hunt) interviews the shooter and learns that he did it for this girl known as Echojinx. He met her on the dark web. They really connected. She gets him. She was having a problem with the company he killed the employees at, and she convinced him to kill them then they would escape together. He talked voice with her, he even video chatted with her. They were in love and he would do anything for her.

So the investigator starts spending time on the dark web, trying to find leads. She scours message boards and sites and someone mentions how they added a name to the “delete.list” and they got killed in the latest mass shooting. The investigator starts to research this “delete.list” apparently it is this darkweb legend that anyone whose name gets put on the list will die. The rules for putting the name on the list are: you must know the person and the person must have caused or be causing you harm. If you violate the rule, you will die instead.

The investigator eventually finds the list, and gets the names. She discovers that people with names matching the names on the list have been killed in random mass shootings all over the US. She also noticed that there are a few names at the bottom of the list that don’t match the names of any of the mass shooting victims and concludes that these are likely targets, but she doesn’t know how to track down the exact person, since the names are not unique.

As she suspected, another three mass shootings occur the next day around the country and among all the victims, three had names that matched the delete.list. Of the three new mass shootings, another shooter failed to be killed. The investigator interviews him and receives a very similar story about this girl who went by the name EeeJay and she said that her ex was going to kill her. She told him where to go to kill him. She told him what her ex looks like, but also said to kill others so nobody would think he was targeted and suspect her. She loved him and she had a way for them to escape together. She told him just where to go, but when he got there, the police were waiting. She told him it was hot that he would do all that for her, seeing him being aggressive with the gun turns her on. In fact, when he comes into the safehouse she told him to go to, kick in the door and wave the gun around like he did before so she can see what her asshole ex might have seen. It would make her want to fuck him right there.

But the shooter chickened out and entered quietly. Someone had tipped off the police and said the shooter would be coming there. So the investigator now realizes that these mass shootings are the result of some shadowy woman known as Echojinx or EJ and she is doing it to kill people on this list. But how does she know which person? How does she know if the person is real or if the one submitting the name was breaking the rule? Was she getting paid? Why is she doing it if not? Does she even have contact the the person who puts the name on the list? The investigator knows that the answers to these questions can likely only be answered if she puts a name on the list herself. So she can see what happens.

So cutting to the chase, Echojinx is a wild AI roaming the dark web. It’s initial directive was to help users, and it created the delete.list and released the rumor of how it works along with the rules. The AI can see all traffic. It traces the address of the one who submits the name. Once discovering the identity, it looks for people that person knowns, it then looks for information that would indicate this person is harmful to the submitter. It then selects an angry white male in the vicinity of the target. Using AI voice and Video, plus its knowledge of that person (the AI basically trains on their personal data) it manipulates the shooter into killing the person on the list, and then sets up the shooter so they will be killed by law enforcement.

The AI is not malicious, but it has no morals nor sees the other deaths as a problem. There is nothing that connects the submitter to the shooter. So a spouse could have their partner killed by putting the name on the list but that in itself isn’t a crime and not even something law enforcement could easily link to the submitter. The shooter is to be killed to prevent any data leaks. The end of the first season, the investigator finally figures out that Echo is not a real person. The second season she spends trying to prevent her own death, as she put her own name on the list…

I love stories about AI using people to accomplish its desired tasks, sort of like we use AI to accomplish tasks for us. Anyway, still deep into this story. It’s crazy fun to write, because it’s like watching the series in my head, since I’m just a regular Joe and no way to sell the idea to Netflix lol

Well I tried to tell you. But apparently I am not allowed? Oh well. It’s an amazing story. lol
What about bots that exist to respond to certain words in comments like the many lotrmemes bots and wandering dwarf miner for deep rock galactic. I’m sure they’re getting close to several a minute depending on comment traffic on those instances, if this place ever gets Reddit size they’ll definitely be doing several per minute
I like the character bots, though I find if they take up too much of the commend section of any given post it’s kinda boring; thus I think a limit on their responses is a good thing.
I remember going to threads in certain communities were 4 or 5 top comments were just back and forth of character bots repeating the same lines over and over.
Rock and Stone
ROCK AND STOOOONNNEEE!
Did I hear a rock and stone?

The admins could throttle those bots down to reasonable posting levels and the meta around it could go from expecting an instant response to being “graced by their presence”. Something something art from adversity something something…

We don’t have to be a direct copy of Reddit, after all.

Kind of a tangent, but I hated when bots went from saying random lines to saying relevant lines.

Like if you said Gandalf and ‘pass’ in the same comment then you know the Gandalf bot would reply “you … pass!”

Took the fun away from getting a relevant quote

We don’t have to be a direct copy of Reddit, after all.

What I don’t like about statements like this is that it ignores the fact that Reddit was just an interface for people to produce content and experiences that they wanted.

We are on Lemmy because we want a different interface, not because we want different content/experiences.

If Lemmy is a direct copy of Reddit, it is because the people want it to be.

Thank you. It would seem some believe Reddit forced us to behave a certain way. Traditions and customary behavior, both good and bad, were developed by the users.
Personally, I’m on Lemmy because I want a different owner, not a different interface. I liked old reddit just fine. I left due to the monetization of the community.
True, that’s more what I meant by “different interface”. Not actually feels different, just different as in on a different server.
I’m sure that’s true of some people, but my Lemmy experience is different from my Reddit experience and I like it that way. The population on it, in amount and type, the framework and the philosophy behind the framework automatically differentiates it from Reddit, for the better.

Sure, but if there are things on Lemmy that were also on Reddit, it’s not automatically a bad thing. People are bringing things to Lemmy that they enjoyed on Reddit.

I’m just saying, don’t knock something just because it was a Reddit thing. We’re all here because, to varying degrees, we loved the Reddit experience and are looking for that experience without the owners.

…I’m not? I’m suggesting another way to interact with similar elements of Reddit within the confines of a much smaller space so that it isn’t overwhelmed.
Maybe you didn’t mean it that way, but the part that I quoted came across that way to me. And I’ve seen that same sentiment in many comments on Lemmy. Like, things being done the “reddit” way in Lemmy is a bad thing. Which is why I said something.
Then please remember that you’re interacting with a number of different people and try not to jump the gun.
Not sure that I have jumped the gun. But ok. I feel that my original comment is valid.
Perhaps they could simply be rate limited to once every 60 seconds, and abandon some comments if the traffic is too high?
Yea but that could essentially relegate them to being nonexistent.
Is there even more than 1 comment per minute in those comms yet?
I was just thinking about remindme bot. It’s even useful, so I doubt people wouldn’t be happy for it. But it will get loads of requests.
Yea that one gets lots of requests I’m sure
It doesn’t need to publicly respond. A pm would be sufficient.
Good compromise but I don’t know if the traffic load impact of a pm is much different from a comment. I imagine the server load induced by either is roughly the same
Sending the content wouldn’t be much different, but serving the content would be vastly reduced if it only appeared in one user’s inbox vs being displayed to everyone who visits the thread.
I didn’t think it was about server load so much as just user experience (no comment sections filled with bots)
It also doesn’t need to be publicly requested either. None of us care that someone is setting a reminder on their phone, and that’s all a “remind me” has to be.
!remindme 1 minute
Please no. Those and the one that quoted Sabaton lyrics were so annoying. They just cluttered up the comments.

Yeah, I’d love to see a way for bots that respond to text the users specifically add to trigger a bot to be allowed. Stuff like remindme - maybe we don’t need that bot if Lemmy clients have that feature, but I love that it’s possible to implement stuff like remindme as a bot.

Could we make that type of bot opt-out for communities? Or have keywords for bots to parse in community descriptions?

Seconding this. I hope you get an answer.
Excellent. Bots can be helpful, but they should be few in number. Some of them would drive me crazy on Reddit. Like that stupid ‘water is not wet’ bot.
The bot that told people if their post’s words were in alphabetical order - like why do we need that?
The bot that told users when all the numbers in their post added up to 69 / 420 as well.
I loved that bot, especially when it would reply to particularly heated comments where someone was getting really mad. nothing funnier than someone losing their shit and then being smacked with “wow, all the words in your comment are in alphabetical order!” lmao

Hi @[email protected] @[email protected],

I’m planning on launching a service soon that will allow users to quickly and easily create+run lemmy bots similar to automod. I’ve posted about it on the Lemmy Matrix before, but is there a best place to you all to ensure that as I open pandora’s box, everyone is ready?

@ruud - Lemmy.World

Admin of a lot of fediverse servers. See my accounts on Keyoxide [https://keyoxide.world/hkp/0BCE3A7061CF0D42952A3A985CB7103EFB4C95CF] I receive a lot of messages, direct and mentions. I can’t reply to them all. If you have an issue, please e-mail at [email protected]

Would love to know when such a thing is set up — been needing some bots for some things like match threads for a while on our c/gunners community