Hi all. I’m trying to create a chatbot that would interact on AOL Instant messenger.

My issue is that i don’t want to create a chabot, have it added to buddy list by many people, only to have AOL block it. Does anyone have experience interacting with AOL on chatbots? I know there’ve been some that had hundreds of members.  Do they officially register with AOL as a chabot? How do you prevent from potential blockage by the service? Anyone knows what thier policies are in that regard? Has anyone tried to ‘register’ thier chatbot with AOL AIM? With any other service? Yahoo messenger? Gtalk? etc?

Thanks! any input appreciated.

Erwin, any thoughts?

I’m afraid I have little experience along these lines, but I wonder… Have you tried contacting AOL with your question? What about looking into their Terms of Service agreement? You could also perhaps contact the other companies with the same questions, I’m sure. Or, someone here may have direct experience with this, and could share their insights.

I can’t find it anymore in AIM’s help system, but for a while AOL had official support for chatbots.

What you’d do is register a screen name for your bot as normal, and then there was a setting you could toggle in that account’s settings on the AIM website, where you say “This screen name is for a bot, not a human” and in that way, you’d register the bot with AOL.

A bot screen name had some special privileges and limitations compared to a regular account, such as:

* They have no rate limit for sending messages, so they can chat with tons of people simultaneously without needing to throttle itself.
* But, bot screen names could not INITIATE new conversations (i.e. “send a message to my friend and say hi” would be impossible if the friend wasn’t already in a conversation with the bot).

I can’t find any information on this on their site though, so I’m not sure if they’ve discontinued this feature.

Back when I ran AIM bots, I’d just have them use normal screen names, and manage the rate limit automatically (I’d program in a 3 second cool-down time after sending an IM to a user, so if nobody was chatting with the bot, it would reply instantaneously, but if a lot of people were chatting it would take longer… and it’d send an “is typing a message…” notification during the waiting periods).

Some other AIM botmasters I knew would implement a mirroring system. They’d sign on multiple screen names at once, and if a user IM’d a bot to start a new conversation, and that screen name was under heavy load (rate limit wise), the bot would IM the user from a different screen name to chat with them there.

Anyway, I haven’t heard of chat service providers having any problems with bots as long as they’re not abusing the system (i.e. spamming) or trying to make a profit from it (i.e. in-message advertising, like what SmarterChild used to do; but SmarterChild had a special relationship with AOL).

And a fun side fact about SmarterChild and related bots: their screen names were technically Admin names, they had NO rate limit, and they were IMMUNE to being blocked or warned (you could try to block them, but they’d never “go offline” when you do). This caused me a bit of stress once when one of my bot’s users told my bot to IM SmarterChild, and I was unable to get my bot to stop because blocking SmarterChild wouldn’t work. I then programmed in a “soft block” system, where the bot would just ignore people without needing to block them at the protocol level.

Along these lines, I’m having trouble with Twitter right now, today:

=> Twitter is not only blocking links to my website, but also my tweets to @Support. What should I do

As I recall, one of the reasons imified.com shut down was because the major IM players were refusing to cooperate.  In light of the new Internet of Things, rumors of the death of XMPP would seem to have been premature.

In the case of Twitter, it is the proactive messaging that they consider to be nuisance spam, more than reactive messaging which is apparently more “kosher”; though, in general Twitter frowns on “@sign” abuse.

Yeah, what I was talking about in my previous post was more for reactive-style bots, i.e. the way SmarterChild worked. When AIM had official support for bot screen names, one of the restrictions was that the bot HAD to be reactive, because the server wouldn’t allow it to initiate conversation with a new person.

Bots that go around initiating all their own conversations would probably be considered spam, even back in the day (AIM/MSN/YMSG bots that were popular circa 2003). Although, a lot of bots would have a feature where somebody could say “send an IM to <friend’s name> and say hi”, and there were some prank services online where you could send a bot at one of your friends to have a 10 minute conversation and let the instigator read the chat log. These would all fall into a gray area though. The prank bots did have an opt-out feature, though.

There are bots on Twitter that are reactive, such as @pixelsorter, you tweet a picture at it and it will pixel sort it and @mention the sender with it. It hasn’t been flagged as a spam bot probably because it’s a reactive-only one (altho there was that one time where two pixel bending bots were tweeting each other back and forth: http://www.theverge.com/2014/9/30/6875163/this-is-what-happens-when-two-pixel-mashing-bots-get-in-a-twitter).