Rogue intelligent agents? Bots have spent years battling on Wikipedia edits
It’s not just humans who wrangle over revising entries on Wikipedia. A new study reveals that automated bots have been engaging in ongoing editing wars on the crowdsourced online encyclopedia for at least a decade.
The new study, undertaken by scientists from the Oxford Internet Institute and the Alan Turing Institute, has revealed that the site’s many editing bots often get into persistent wars wherein one bot changes the edit the other bot has done – and back and forth.
Wiki’s bots, which made around 15 percent of the edits on the site in 2014, are employed for all kids of tasks on its 41 million-plus pages. This might be cleaning-up of an act of vandalism, adding links, checking spelling or copyright infringement, but also modifying the content of articles.
According to the study, the more bots that were employed, the deeper the editing wrangles became, sometimes leading to bots being taken out of action. The problem is, while bots have to be approved according to the site’s bot policy, the study says they are then left to some extent unchecked “without a formal mechanism for coordination with other bot owners.”
The study focused on 13 different language editions of Wikipedia and edits over a period of 10 years from 2001 to 2011. Dr. Milena Tsvetkova from the Oxford Internet Institute not surprisingly said bots behave differently than human editors, but their sometimes strange behavior was a concern.
Language played a part in which pages were contested the most, with the German-language pages seeing the least editing-wars – only 24 instances of bots revising each other’s work in a decade. The English language version saw bots fighting over edits 105 times in the same time frame and the Portuguese version 185 times.
The study also found that some diverse pages were hotspots of rivalry, some of which were the pages of former president of Pakistan Pervez Musharraf, the Arabic language, Uzbekistan, Estonia, Belarus, Niels Bohr and Arnold Schwarzenegger.
Some bot battles were worse than others, with Xqbot and Darknessbot having the most disagreements on a number of subjects from English football clubs to Alexander of Greece. It was found that the former reverted Darknessbot’s edits 2,000 times, while Darknessbot changed Xqbot’s editing 1,700 times.
“This has implications not only for how we design artificial agents but also for how we study them,” said Tsvetkova. “We need more research into the sociology of bots.”
Although the study admits the bot wars were small in comparison to the good work the bots are doing, the continuous disagreements were seen to be “inefficient as a waste of resources, and inefficacious, for it may lead to local impasse.”
The researchers said the findings bring attention to how artificially intelligent agents are designed, concluding, “As bots continue to proliferate and become more sophisticated, social scientists will need to devote more attention to understanding their culture and social life.”
Photo: Zen Skillicorn via Flickr
A message from John Furrier, co-founder of SiliconANGLE:
Your vote of support is important to us and it helps us keep the content FREE.
One click below supports our mission to provide free, deep, and relevant content.
Join our community on YouTube
Join the community that includes more than 15,000 #CubeAlumni experts, including Amazon.com CEO Andy Jassy, Dell Technologies founder and CEO Michael Dell, Intel CEO Pat Gelsinger, and many more luminaries and experts.
THANK YOU