Wikipedia Creates AI System To Filter Out Bad Edits (thestack.com)
An anonymous reader writes: Wikipedia has developed a new artificial intelligence system aimed at improving the quality of its entries and detecting both mistakes and damaging edits made to its articles. The technology is named the Objective Revision Evaluation Service. The Wikimedia blog explains that the system is able to highlight incorrect edits, allowing editors to filter them out from the "torrent" of new amends and scrutinize their credibility. The entire service and process is open – with Wikipedia making revision scoring transparent and audit-able by publishing the source code, performance statistics and project documentation publicly under open licenses.
Now if the AI can be programmed to weed out the;
*Man-Child admins who lack authority in real life and become a Wikipedia admin to (over)compensate
*Agenda Warriors who go there to further their ideological battles
*Bullies
*Harassers
*Stalkers
*Jimbo Wales
we might see some improvement.
Why a joke? Consider it a spam filter for Wikipedia entries. The more popular entries have unfortunately had to get locked down due to rampant vandalism. Wikipedia, for all it's faults, is one of the wonders of the modern age. Between it and Google, it's like a global repository of human knowledge... or at least, a summary of human knowledge, with links to deeper knowledge.
But while its power is derived from the masses of humans that create and edit that content, the masses are also its weakness. Humans are endlessly creative, and this tool will probably just teach trolls how to vandalize more creatively in order to fool the algorithm. Still, if it cuts down the noise for editors, it may end up being a good thing. It's hard to say until it's actually been deployed in the real world for a while.
Irony: Agile development has too much intertia to be abandoned now.