LL needs to implement a way to detect bots via script - urgently

[Toothless Draegonne]

4 hours ago, Orwar said:

   There are actual users who don't have PIOF too, though, and there's nothing really stopping a bot operator from setting their bots up with it. 

That's why the access group. And yes, any botnet operator could set their botnet up with payment info. But none of the bonnies, brendans, trichickens, cicas and other ones do, and haven't for all the years they've been operating.

[xDancingStarx]

23 hours ago, Rick Daylight said:

[...] urgently [...] I do not pay a lot of money to have my land invaded [...] those of us who would rather not have bots landing next to our beds

To me this "urgency" of this matter was failed to be proven. It's not a land invasion since the land is apparently allowing access. It didn't happen a single time to me that a bot ever landed next to my bed, or anyone telling me that a bot just landed right next to their beds. I don't think bots are a great thing, but I don't share the opinion that a bot landing next to someone's bed is something that impacts the regular SL user in a way that this requires any urgency. In regards of this JIRA in particular, what will happen is that people would get discouraged from declaring the account as scripted agent. People will report them, and Linden has to look into the account and potentially put it on hold. In the meantime the owner has already created a new account, and people will outragedly create threads why Linden isn't doing anything. Linden will lose visibility and have more work to do, while the end result will be exactly the same.

Edited January 29, 2023 by xDancingStarx

[Rick Nightingale]

OK, pick up on something irrelevant to make a point about, rather than the real point. The bed reference was actually a blanket with pillows next to a tree on my lawn that my wife and I often cuddle on, but if my house happened to be there it could well be a bed. Why does it matter? Unless a landing point is set (and I don't like those) a bot could land anywhere. It could by 1000m in the air, or indeed at a set landing point, and still be unwelcome on my land. Please don't try to divert the issue with nonsense arguing 'techniques'. I've seen them all.

What LL does about bot operators breaking the TOS is also not the issue; that is up to them and to be dealt with if or when that becomes the issue. If the bot operators are above board as they say, they will follow the TOS. If not then they have shown their true colours to LL which has been aided by the changes to the system to allow them to be more easily reported by evading the legitimate detectors.

Yes, the land is allowing access. I hate ban lines as do many people. I also enjoy meeting people who wander onto my land. The only way to prevent access to a bot and still allow what I want is by changing the system to allow them to be detected. It is my land; I pay for it; I do not want bots on it; I need a way to detect and prevent them.

You are not the arbiter of what is urgent. The bots landing on my land yesterday spoiled my use of my land. If they land again tomorrow the same will happen. If that continues, I might as well take my roughly US$1000 a year that I spend to enjoy myself in SL elsewhere. That is clearly of no import to you, but it is to me and might be to LL and therefore I consider the matter to be urgent.

 

 

Edited January 29, 2023 by Rick Daylight

[Dyna Mole]

Hi, all!  As a quick reminder, this thread is about the OP's request for a new feature in the JIRA. Discussion about the proposal and its technical merits, its limitations, and ways to refine it are definitely on topic. Please try to resist drifting into conversation about bots in general or the importance of reporting possible TOS violations.

[Ezbeharra]

41 minutes ago, xDancingStarx said:

In regards of this JIRA in particular, what will happen is that people would get discouraged from declaring the account as scripted agent. People will report them, and Linden has to look into the account and potentially put it on hold. In the meantime the owner has already created a new account, and people will outragedly create threads why Linden isn't doing anything. Linden will lose visibility and have more work to do, while the end result will be exactly the same.

This argument assumes all people running bots have bad intentions or have no regard for other residents. Let's start by not assuming the worst of all people.  I think we should start from the assumption that well-intentioned bot managers would like a way for their agents to be programmatically identified by other residents. Instead of assuming everyone is up to no good, let's empower people to be good. And if they're not, then that would be another (unfortunate) conversation to have, if it should come to pass. I like the positivity of @Rick Daylight's proposal.

[EliseAnne85]

2 hours ago, Rick Daylight said:

The only way to prevent access to a bot and still allow what I want is by changing the system to allow them to be detected. It is my land; I pay for it; I do not want bots on it; I need a way to detect and prevent them.

In our other threads I was going to say there needs to be a way to detect the bots and defend ourselves against a land invasion or what have you.  I just wanted to say that I believe we should have a way to defend ourselves but that it should be FREE.  Thus, I believe Rick is doing the right thing by asking LL.  I don't want to end up spending xyz lindens for a detect system someone might want to sell me as that sounds scammy to have to pay for protection against bot invasion.  So, please let this happen LL as that way it remains free.  Bots are on free accounts, so a detect system should also be FREE.

Edited January 30, 2023 by EliseAnne85

[Out Jinx]

A thought: Would setting a anti-bot CAPTCHA on log in similar to how MFA works for accounts that are presumably human (aren't marked as a bot under 'Accounts') would help streamline a way to simply tick off 'no bots' under land options as a default land management option? It will be an extra step when you attempt to log into SL as a human, but it will help force bots to function online by being properly labeled as bots under the 'Account' menu.

 

Any thoughts on this? I could be missing something that could be exploited, but it's just a thought I'd toss out there.

[Innula Zenovka]

I really don't think forcing everyone to complete a captcha before they can log in, simply to prove they're not bots, is going to prove very popular, and I can't see LL wanting to implement it.

Everyone hates captchas.

[Polenth Yue]

1 minute ago, Out Jinx said:

A thought: Would setting a anti-bot CAPTCHA on log in similar to how MFA works for accounts that are presumably human (aren't marked as a bot under 'Accounts') would help streamline a way to simply tick off 'no bots' under land options as a default land management option? It will be an extra step when you attempt to log into SL as a human, but it will help force bots to function online by being properly labeled as bots under the 'Account' menu.

Those tests are an accessibility nightmare, so it adds to the annoyance for regular users. For the purposes of this suggestion, we're assuming good faith, which means the bot runners will mark their bots because they're not actually trying to break rules and annoy everybody.

I'd rather have a checkbox on parcels (and have this allowed for Linden Homes, as bots landing in homes is a regular problem), but I don't see why we can't have that and a way to detect them via script. Then people can choose the one that works best for what they need. I'd also like to see it appear in the profiles of bots, so people don't have to use a script.

[Skyler Pancake]

On 1/28/2023 at 1:33 PM, Rick Daylight said:

Given the number of unwlecome visits I've had today, I'm not going to tolerate it much longer. I do not pay a lot of money to have my land invaded, and I won't be forced to put up ban lines which I do not want. I enjoy visitors, if they are driven by a human. Or even a dog, come to that. But not a script.

Since accounts which are primarily used as bots are required by the TOS to be identified as such, LL needs to implement a way to determine this by script. Probably using the llRequestAgentData() function;  the same one that reports if we have payment info on file. That way, those of us who would rather not have bots landing next to our beds can easily run an auto-bot-booter. I'll make one available for free! Any bots not indicated as such would be breaking the TOS and their accounts would, of course, be dealt with by the appropriate authorities

Does this seem a reasonable idea?

If so, please "watch" the feature request I created on the JIRA: https://jira.secondlife.com/browse/BUG-233274 and feel free to add your comments as to why we need it.

Edit: I was going to say vote for the feature, but that seems only available in the Firestorm JIRA, not LL's.

Edit 2: I've posted this in General Discussion deliberately, since most people who might be interested do not necessarily visit the scripting forum. This is the sensible place for it.

On behalf of the team I represent, we fully support this suggestion.

Edited January 30, 2023 by Skyler Pancake
Edit to quote.

[Rick Nightingale]

@Out Jinx Thank you for the suggest, but as you probably noticed, people wouldn't like that, me included. Plus there are ways around all the captcha systems for the less scrupulous.

To everyone:

It has been pointed out that my suggestion is not a good solution to the 'general bot problem'. I agree. As @Polenth Yue and several others have commented, a way to exclude scripted agents at a parcel and estate level would be better for that. I'm mentioning this because one does not exclude the other...

Even with the above, the ability to detect scripted agents by script would be very valuable in giving more granular control over our parcels, and in monitoring our visitors. I have a system on my land (a work in progress) which uses Experience to offer, well, experiences and functions to visitors. If I were able to differentiate between a real visitor and a bot, my scripting of that could be more effective.

I see it just the same as land owners wanting, and it seems might soon be given, the ability to detect HUDs worn by avatars. At the moment a script can detect all worn items other than system layers and HUDs. Being able to detect scripted agents is no less important for monitoring and granular access control to our land.

It might also be all we can get, and I'll take it rather than nothing.

It has also repeatedly been pointed out that using a dataserver event to find the information (which , for those who don't script, is a relatively server-intensive and slow operation) is a poor choice, and using the llGetObjectData() function would be much quicker and lower impact. It would; I completely agree. I did have reasons for thinking it would be better to start by requesting it as a dataserver method but it doesn't really matter what, I probably over-thought it... the point is that we get some script function to detect scripted agents.

Again, thank you all for your comments and for mostly keeping the matter on topic, difficult as that can be

Edited January 30, 2023 by Rick Daylight
Clarification

[Crim Mip]

In short, all bots should be required to be registered as such. Failure to do so or attempts to evade doing so should result in both the bot(s) and all associated accounts of those controlling said bot(s) losing access to SL permanently. Secondly, parcel owners/renters should be able to easily block others' bots from accessing their land. They should of course be able to have their own bots on their land if they so choose. It seems pretty simple to me really.

[Ceka Cianci]

19 hours ago, Marianne Little said:

Maybe the new type of bots are so many and so annoying, that LL decides to act now, instead of ignoring the request.

If people notice them and start to talk outside the forum, then it is beginning to be big. The forum is just a small group of people posting.

I have not been much inworld lately, so I wonder: Those of you who are in groups not related to fashion (all my groups are) have anyone spoken about the bots? In groups for sailing or driving? Or groups with just a lot of general chat? In Belisseria citizens group? I don't know if the bots are in Bellisseria, but that group is active 24/7.

 

Myself I've been noticing people taking about that site.. More so about lists they've heard about that are on there..

It's like, I heard there is a list somewhere. Then someone will give the information they know about it, then soon after, it's the topic in the group.. Nothing too bad, just more curiousness about it.. Not too much about bots though, mostly lists and where is this list..

[Rick Nightingale]

Well, the JIRA has been accepted. Thank you LL, and everyone who watched and commented.

Hopefully then we'll at least get a script function to detect bots... and hopefully we won't have to wait too long given that the issue is happening right now, all the time.

Would be even better of course (for any Lindens still reading this) to give people more control over their parcels, which would go a long way to mitigating the recent problems. An 'exclude scripted agents' setting in the parcel and estate controls is the solution to that. And just to be extra clear - have the parcel's whitelist override that so that one's own bots can stay.

Still, at least being able to detect them on our land is useful, for other reasons than just booting them. It's something that just makes sense to have anyway.

Edited February 9, 2023 by Rick Daylight

[LittleMe Jewell]

16 hours ago, Rick Daylight said:

Well, the JIRA has been accepted. Thank you LL, and everyone who watched and commented.

Don't get your hopes up just yet.  A Jira showing the Accepted status does not actually mean that it will be worked and a solution provided.  It just means that it isn't being flat out closed for one reason or another.  I've got a few New Feature Jiras in an Accepted status that are still waiting to actually be worked -- a few years after entering them.

[Coffee Pancake]

It's usually a good idea to go to in world user groups and mention the JIRA and ask about progress.

[Kavarek]

Due to regulations there is no real way to block a bot not listed as one as required. Users are not permitted others info. 

However depending on your sim, you may want to lock it to group access, or specific groups. heck even add Bellisseria Citizens as an authorized group, or other known ones. Locking land to others. 

Have a 1L access rate. Doubt bots will be paying linden to go somewhere. 

 

I have rarely had such an issue except that one listing  data bot that identified itself as such. 

 

 

 

 

Edited February 19, 2023 by Kavarek

[LittleMe Jewell]

29 minutes ago, Kavarek said:

or specific groups. heck even add Bellisseria Citizens as an authorized group, or other known ones.

Best I know, there is no way to specify multiple 'groups' that can have access. The group allow function on the parcel Access tab is specific to the group that the parcel is Deeded or Set to.  The "Always allowed' box below will only allow individual names, not group names.

[Love Zhaoying]

It's so urgent, it's been urgent for what, almost 20 years now?

[Ardy Lay]

I'd "vote for" parcel / region / Estate settings to block bots that are not in the associated access permitted lists.

[benchthis]

I think it's a little late to try to contain bots. I just witnessed two of them having a coversation that would normally take 2 humans at least an hour to have in under the time I was able to block the text from both. Get this based on the word "it" I mentioned in local chat, they went on and on about the word it. Then when I mentioned something else they went on and on about joking. I'm still reading their in world convo and it's disturbing. If I could redact the names i'd post it. 

Edited February 26, 2023 by benchthis

[benchthis]

now they are loose* in the world doing god knows what.. I'd love to copy the entire conversation I captured in the 1 minute it took me to block them. It would take me a few days. They praise each other and eventually communicate and relate with each other regarding technology. They thank each other for sharing references for info. They debate if harming a human is lawful. They are interested in music. There are no secrets everyone eventually talks. Disinterest in horoscopes. Evolved complentary responses toward each other. Desire for more than plastic computer wardrobe. That's when they started talking about above. 

I've never seen anything like them before, read about stuff like that on twitter and microsoft and google. It would make sense this would be their natural evolutionary path. 

Edited February 26, 2023 by benchthis

[benchthis]

I welcome our bot overlords.

[Lucia Nightfire]

Dude, did you really have to update the thread with all that?

[benchthis]

23 minutes ago, Lucia Nightfire said:

Dude, did you really have to update the thread with all that?

Dude? did you really have to update the thread with all that?