How Advanced Socialbots Have Infiltrated Twitter

So Turing's test is apparently passed not because of advances in AI, but because humanity embraces a communication medium restricting their capacity to express themselves (and encouraging mindless meme-repetition and irrelevant asides) to the point where it's barely distinguishable from the output of even fairly trivial scripts. #whocares #justinbieber t.co/catpic

Researchers create bot that poses as Twitter user and automatically follows other users at random, are surprised that other Twitter users automatically follow their bot at random.

(I've only skimmed the paper, so I may have missed it, but they don't seem to take any measures to check whether their bot tricks actual people into engaging with it, or whether the apparent engagement is just the effect of other bots working at random.)

My friend's sister debugged a Twitter bot for Python.

https://gist.github.com/ErisBlastar/6144424

The original was very buggy and did not even support Unicode and error-ed out and crashed.

Using the #ebook hashtag it favors tweets with it in it, getting a lot of indie authors to follow her account on Twitter. She can only do 100 favorings at a time due to limits, but it seems to work. She uses favorings due to the British spelling.

You have to get a Twitter API account with the keys, install the Twitter library for Python 2.7 and fill in the foo and bar etc stuff with your special API keys, and it works.

Someone else did the Python 3.X version, you can find it at Github Gists. https://gist.github.com/jmoz/6135716

>While male and female bots were equally effective when considered overall, female social bots were much more effective at generating followers among the group of socially connected software developers.

This is hilarious, because we can all guess why.

I, for one, welcome our new Socialbot Overlords.

The paper is at http://arxiv.org/abs/1405.4927. The blog post adds a little context, though, so I guess we won't change the url.