The tag system in Twitter is an effective way in distributing information beyond a user's social network. However, the system can easily be contaminated by spam messages. In western countries, most of the spammers are driven by commercial interest, while in China, they are political tools for disrupting information flow.
Frustrated by political spam, Chinese blogger Jiruan looks into the pattern of Chinese political spammers or the so-called 50 cent party by analyzing the tag #aiww which represents prominent artist-activist Ai Weiwei who have been arrested since early April 2011.
The chart below shows 8000 tweets from April 26 2011 to May 2 2011 based on the search term “#aiww”.
The tweets come from 26 users with 22.51% coming from user @lidamink and 22.43% coming from user @eiuielk. Whenever someone searches #aiww in Twitter, about 80% of the messages are either spam or defamatory remarks. Among the 7 search result below, 4 messages came from @lidamink and @eiuielk:
The blogger also analyses the top 10 users of the #aiww tag and found out that all of them are spammers and their tweets have made up to 75.06% spam messages in the #aiww search term:
1. @lidamink 22.51%
2. @eiuielk 22.43%
3. @ideletey 6.20%
4. @jchang_22079 5.8%
5. @xiazuhu 3.38%
6. @andingzirua 3.27% Account Suspended
7. @lalivea 发帖3.09%
8. @liunianzhijian 发帖2.87% Account Suspended
9. @nizhexin 发帖2.85% Account Suspended
10. @bichunlong 发帖2.66% Account Suspended
Some Twitter users have compiled their own list of political spammers or 50 cents to alert other users. The account @eiuielk has been filed in 6 Wumao (50 cents) lists:
Apart from #aiww, other political tags such as #cnjasmine (aggregator of Chinese Jasmine protest information) #5mao (aggregator of 50 cent party) are also contaminated by political spammers. Jiruan urges Twitter user to actively report spam messages.
Twitter needs to implement hashtag controls and to delegate more control to its users. If this was IRC, a forum, a BBS, or a blog, the users could write scripts to ban spammers automatically.