When will Yahoo end the use of ‘bots and automated scripts in its chatrooms?
I used to really enjoy using Yahoo Chat, but unfortunately, as is expected in a free no-charge resource, you always have a few no-goodniks ruining the party for everyone else. Oddly enough, the answer to ending the user of bots and automated chat scripts used to hawk everything from dating sites to ripoff webcam sites is already implemented in other sections such as Yahoo Personals and Yahoo Photos. In those sections you are required to repeat alphanumeric characters shown in a small graphic in order to use the services. Sadly though, the one section of Yahoo where this is most needed is the one place where it is not implemented–Yahoo Chat. If this oversight continues, soon there will no real people using the Chat section because it will be so overrun with bots and scripts as to be unusable to the people it was originally designed for — Yahoo HUMAN users themselves.
Yahoo invented this application, so I know they can come up with a way to keep automated scripts from abusing it. Every other application has, and they’re nowhere near as capable as Yahoo Messenger. What Yahoo needs to do is stop making excuses and fix the blasted thing!
April 14th, 2010 at 10:07 am
Unfortunately, the “bots” as they’re called, connect to the chat rooms (channels) the same way normal users do, so there is no automated way of filtering them out. Even the pages that require “retyping” as you described can be deciphered by intelligent scripts, and are in no way foolproof.
The only way Yahoo! could manage the chat rooms would be to hire room administrators to manually filter users. Volunteers could be used, but on a portal so large as Yahoo!, is is unwise to give such overwhelming power to a few people.
Report this comment
April 14th, 2010 at 11:11 am
As a response to the person claiming the only solution would be to have administrators, that’s just not the case. For one, the tests, or CAPTCHA, which the users are required to fill out, vary in strength. While it is true that the one which Yahoo uses is volunerable, it’s not the case that they all necessarily must be.
Also, yahoo could begin to use something like what eBay uses, where all members monitor all other members. Yes the bots could report real people as bots, but chances are good that with some fine tuning, you could get a consitent algorithm which usually works, similar to how Google page rankings work. And similar to the way Google must constantly update their logic when people find holes in it and unfairly rank their websites, the logic on Yahoo would need to be fine tuned occasionally.
For example, I’d suggest some scoring based on the way that spam is scored, so there would be some component of different words in the links, the randomness of the message, repeating a message, etc. And if the score were too high, all you would need to do to reactivate is answer a tricky CAPTCHA (possibly with something beyond an image component). Even better is varying the type of CAPTCHA so it’s very hard to tune the breaking algorithm for it. Or how about adding an audio component?
Realistically, when a profile has 3 websites linked in it, has a certain occurance of words in it, and that same user posts links very very often in chat, it’s probably a bot. If everyone flags it as a bot using the “report as spam” function, you can be pretty sure.
So no, there is no easy solution. However, I believe with proper engineering there is a solution which would take care of the majority of the problem. And if you make it hard enough, the spammers will move elsewhere.
http://www.cs.sfu.ca/~mori/research/gimpy/ breaking yahoo’s captcha, and more info on the subject
Report this comment