Wikipedia has announced it will be rolling out an AI engine to help identify badly edited content and vandals who maliciously edit articles, in an attempt to police the open-source project.
The machine-learning tools - called Objective Revision Evaluation Service (ORES) - have been collated by the community's researchers and specialists and will go some way to automating the editing process and ensuring content is always fair and trustworthy.
Additionally, it will mean more articles can be vetted, then deleted or updated if they are found to be malicious or incorrect.
ORES uses APIs and trains models against edit- and article-quality assessments made by Wikipedia editors, commonly known as Wikipedians. It then creates a score based upon these individual edits and articles as a whole.
The APIs test using various different parameters, such as whether the information could be damaging, and outputs the response in JSON format. The scores are produced in a median timescale of 50 milliseconds for already scored revisions to 100 milliseconds for un-assessed revisions and can be calculated across 14 of the 20 languages supported by Wikipedia.
"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," Wikipedia said.
"Our hope is that ORES will enable critical advancements in how we do quality controlchanges that will both make quality control work more efficient and make Wikipedia a more welcoming place for new editors."
Wikipedia has been testing the service with more than a dozen editing tools and it says it's already beating the state of the art systems in the accuracy of its predictions.
The next stages of ORES's development are supporting more wikis, automatic categorisation of edits and bias detection.
