One option that strikes a compromise between blocking all messages from new users and allowing new users to message anyone and everyone else on the site: what about new users can only message curators and/or admins who volunteer as “new user mentors” until these new users have done enough legitimate use of the site to show that they are actual users and not spammers? This way, the people who would receive the spam messages are people who could flag and suspend the accounts, and the regular users would be none the wiser.
Better be careful in this arms evolution, the spammer is an animal that just becomes more sophisticated the harder the criteria become. eg If we require they post observations etc, they will develop a bot to post observations, and IDs, and we will end up with lots of fake observations and IDs being posted …
Ah, but do we tell them that’s what they have to do? ;)
Maybe every new account has an arbitrary, randomly generated number of common user tasks assigned to it that a real human would do in the course of having a new account, but that a spammer wouldn’t bother wasting their time on.
make a new user identify to coarse taxonomic units (plantae, etc) and maybe we can ‘breed’ the spam bots to do that task for us
I agree with @charlie. Like some CAPTCHA methods if there were some way you could have a grid of nine “unknown” observations and the user/bot had to select all observations that had an obvious subject that was at least life. Maybe it could at least cull the better or identifiable “unknown” observations.
It could be part of a tutorial like @graysquirrel mentioned, perhaps on what are good examples of observation photos. Maybe somehow make it fun for actual users and more complicated for the bots but if they actually made it through it at least provided some benefit for iNaturalist.
Yes, great lateral thinking. I suspect that the bot writers just out-source their captcha identifications - ie they have a bot that needs to create accounts on some captcha protected site, all they need is their own website offering something ‘free’ that people search for all the time (eg windows drivers, manuals, or that other graphic commodity that pervades the internet) - then each time someone comes to their site the bot starts to create an account on their target site, grabs the ‘captcha’ challenge and presents it to their unwitting dupe who happily solves the captcha for their bot … ever wonder why some sites keep challenging you with captcha even though you think you’ve clicked correctly ;-)
So, yes get their bots to find unwitting dupes to identify ‘unkown’ observations at least down to plant/animal :-)
off topic, but i think iNaturalist should develop its own CAPTCHA (if it hasn’t already). you could take a random research grade observation and throw it into a modified Compare tool that pulls in the community IDed species, a top suggestion from computer vision that’s not the Community ID, plus 2 other random species in the same class, and ask the new person to pick the right species. that could familiarize new people with the compare tool and also provide data that could help spot problems in IDs and computer vision.
EDIT: I just got to @briangooding’s CAPTCHA suggestion, and i like that, too.
I also use messages a lot, specifically to welcome new users who are in my project, troubleshoot problems with them, etc. Perhaps it could be good to allow an individual user to turn off the ability for anyone to message her or him, but not to remove messages in general. In fact I would like some form of messages to be visible on the app, at least as a notification directing the app user to check the desktop site.