"You are chatting with a bot" forced IM from scripted agents
Glaznah Gassner
Scripted agents recently started sending a warning message to the residents:
"You are chatting with a bot, companybro Resident. Do not share any personal information. Learn more at https://second.life/scripted-agents."
This causes an extreme inconvenience because SL business owners are using scripted agents (bots) for automated communication: in-world support, automated notifications, greeters.
Moreover, this IM repeats over and over for the same agents+residents.
The automated IM ruins all the user experience, especially for entertainment and business cases. While the goals of the IM is clear, can this feature be disabled/postponed for now to be discussed?
Log In
Beatrice Voxel
I've been getting that on an alt I leave logged in for spotting duty at SML gyms. There is no 'scripted agent' tag, we're using RLV plus an attachment for gym visitors to position the spotters on the appropriate equipment. Yet SOMETHING has flagged that account as a bot? I do have an "unavailable" chat autoresponse set from the viewer's settings, stating which gym to go to for spotter help, but that's it. No idea how to even fix this problem.
VriSeriphim Resident
Pre- or post-fixing the agent name with BOT seems reasonable.
Showing warning as a system message ONCE per session seems reasonable.
Any more than that seems excessive.
Devilgrey Resident
maybe it just needs to send this notice once at the start of an interaction?more than that seems a little much
Darcy Linden
Hi Giaznah, we're having discussions around this topic and realized there is a fine line between UX and privacy protection. We're tracking this issue and thank you for the ticket submission!
Glaznah Gassner
Darcy Linden Thank you for responding! The direct IM response is just too invasive.
However, I just realised that it is a good way for _bots_ to detect other bots. We have some autoresponders tools, and it is a pain to prevent dead-loops when bots are talking to each other.
This notification IM is a perfect way to stop autoresponding. I just wish it to look somewhat different.
Do viewer creators (including me) have any ability to detect this is a system message (besides comparing the text)?
Grimm Reaper
i second this, i want it removed because if i want to have a relationship with the bot i dont need this
JenniWindrider Resident
If bots weren't abused by a few, this message wouldn't be needed. The problem is some bumwipes using AI bots to deceive regular residents.
Although I do agree it's invasive, I think it's currently a necessity. Even if it's a Customer Service bot, I absolutely want to know at a glance if that's a bot or a person. WITHOUT having to open the profile to check for Scripted Agent status.
Devilgrey Resident
JenniWindrider Resident agree. a label next to the the username kinda like grouptags would be beneficial. with ai-npcs/companions getting pushed into sl, bots need to be clearly identifiable inworld and in IMs
hacker Resident
I think a better option would be to add a BOT tag to the name and profile, similar to how it's done on Discord, Twitter, etc. and when hovering over the tag, it would display this message
Devilgrey Resident
hacker Resident yeah i think bot tags needs to be in both places because i dont want to check profiles of everyone to see if its a bot or not
Devilgrey Resident
hacker Resident maybe this should get it's own feedback thread? because honestly automatic bot tags for scripted agents would help stop ppl using bots to deceive others, and also make it clear who's a bot vs human.
Chaser Zaks
It'll be a more viewer side feature in the future:
The simulator now sends a warning message when receiving an IM from a scripted agent
* The message appears once per viewer session, or every 24 hours (whichever is more frequent)
* A viewer currently in development presents the warning properly as a system message
Glaznah Gassner
Chaser Zaks Thank you, I missed that. Repeating this warning every viewer session is very annoying for the users.
Why don't LL wait for development viewer to get released then...
Rachelle Kiyori
LL could have handled this way differently. A message in the agent's profile ("This is a scripted agent") would have been just fine. This is too much going too far.