I have a rather simplistic solution for acceptable bot behavior as to making comments, which is the most annoying form of bot spam. Now, there is nothing that forces a bot writer to conform to this, but i think a lot of bot writers want to know what the rules are and if they know what is and isnt acceptable, it will at least keep the honest bot writers from being downvoted.
The idea is that if a bot follows these rules, the community wont downvote it into invisibility so by following the rules, it is better for the bot writer. The author of the bot would be able to describe his bot's functions as appropriate with additional posts, but that is subject to posting conventions.
I propose the following for an acceptable bot comment behavior:
- introduceyourself post is made on behalf of the bot that describes what it does and how to invoke it
- bot follows what is described and only generates comments when invoked by a user via @botname
- spoofing a user request so as to be able to post a comment is not allowed, ie. "bot" would include any and all bots that are interacting
- bots that conform to this wont be downvoted by community and if the response is useful it is nice to give an upvote
For example, there is the isaac.asimov bot that posts about reading levels, which is annoying when it is done on every post. However, I could see how it could be useful for some people some of the time. So proper behavior is for isaac.asimov to monitor posts and if and only if it sees @isaac.asimov would it post a comment.
Now parsing posts is not so hard if you are just looking for "@isaac.asimov", but to do anything more complex could be an issue, so whether a bot can take additional parameters or not would be described in the introduceyourself.
Based on this, we will get a set of useful bots that need to compete on the functionality they offer, ie. bots that dont do anything useful wont ever be invoked. Also to show appreciation for the bot, if the answer they give when invoked was useful, upvote it. This way the bot writer's have a clear path to revenue, eg. make a useful bot and its responses will get upvoted.
Let me know what you think about this proposal or just upvote it if you like it.
James
crypto777 is my witness account #37 on https://steemit.com/~witnesses
Interesting!
So a bot would only comment if the person writing the post asked it to, by inserting its name into the post? That might be helpful, if bot creators follow that guideline.
Could you elaborate more on #3, not sure i understand that.
Given the rule that if a user invokes a bot with @botname and that is what is required for the bot to post, what if there was a rogue bot that posts @botname? Then the "innocent" bot would respond and be following the rules.
There are thus 2 scenarios, eg. the rogue bot is run by the same person that runs @botname, in which case it is a violation. The other case would be that some anti-bot bot makes this comment to make it look like @botname is breaking the rules. So, we probably need some way to resolve this or maybe @botname needs to have a blacklist, but we are now getting two levels deep and I think we can deal with such issues after the primary one of botspam is dealt with
Thanks for explaining this. I think a blacklist might work, and it might be something that the bot creating community could do collectively and share. I'm sure there might be other solutions that will be tested as well, when any issues arise.
Good idea to promote people makeing bots that actually help users... this however might be useful for bot developers that actually want to help the community.
The only problem I see is that most bots are made with the idea to make profit for their authors by upvoting content or doing somethign else, so may of these probably would not want to introduce their bots let alone make them do something only if invoked by a user.
that is why it is important for the community to upvote the bot's useful responses. If the bot writer's can make more money from making useful responses, they will do that instead of spamming.
Basically instead of a neverending war (like virus vs antivirus), steemit humans can reward good behaving bots and thus incentivize the whitebots. As soon as the bot writers see there is more money making whitebots than blackbots, instead of an unwinnable war we end up with proliferation of useful bots. At least that's the theory...
of course there will always be spam bots that nobody likes, but by having a clear path to profits for botwriters, it can only reduce the spam bots
Check the Hidden comments, there is an interesting comment from a bot that linkes to an article by a bot operator... a major BOT WHALE apparently :)
I could also imagine Steemit.com eventually building a sidebar where bots could compete for visualization such as displaying the reading level, "cheater score" and other qualities. Could upvote as well to see which bots got top billing on each post or the bots would take up the top few spots based on their average monthly upvotes/reputation.
I believe we shouldn't use bots at all