Wikipedia launches artificially intelligent editor to filter false entries
Wikipedia has launched a new artificially intelligent tool to highlight new edits deemed to be 'low quality' in an effort to help its human editors to quickly spot suspicious or false changes to articles.
The website, which can be edited by anyone and sees 10 edits made every second, said the Objective Revision Evaluation Service (ORES) has been designed to act "like a pair of X-ray specs." Combining open data and open source machine learning algorithms, the editor can spot article edits likely to be low quality based on the language used and context.
These edits are then flagged up for a human editor who will make the final decision.
Key figures to fall foul of Wikipedia's open editing policy include Bill Gates, whose picture was changed to a include a crudely drawn pair of devil horns and moustache, Google's Sergey Brin, who pranksters claimed was in a relationship with Wikipedia founder Jimmy Wales in 2006, while another edited Tony Blair's page to state he hung "posters of Adolf Hitler on his bedroom wall as a teenager."
"By combining open data and open source machine learning algorithms, our goal is to make quality control in Wikipedia more transparent, auditable, and easy to experiment with," Aaron Halfaker and Dario Taraborelli of the Wikimedia Foundation wrote in a blog post.
The software will encourage more first time editors, known as Wikipedians, to join up as prior damage-detection tools have encouraged the wholesale rejection of new editors' changes under the impression they were vandalising rather than contributing to articles in good faith.
Wikipedia is the internet's largest online encyclopedia, with more than five million articles covering less than five per cent of humankind's knowledge.
Last month WIkipedia founder Jimmy Wales, who is a long-term defender of privacy rights, described the Investigatory Powers Bill (also known as the Snooper's Charter) as "stupid", and implied that the industry should take a stand, suggesting Apple should stop selling iPhones in the UK should the government ban end-to-end encryption.