CAMBRIDGE, Mass. — Wikipedia, the world’s free online encyclopedia, is a true wonder of the internet. As vast a resource it is on quite literally any subject one can imagine, it’s knock is that it allows anyone to make edits and provide updates. That means sometimes articles show outdated or inaccurate information. Misinformation feels like it’s reaching a fever pitch in 2020, and in an effort to create a more accurate internet, researchers at MIT have developed an advanced AI system capable of automatically identifying and correcting inaccurate, outdated information presented in Wikipedia articles.
For now, Wikipedia articles are painstakingly reviewed and slowly corrected thanks to the efforts of human volunteers all over the world. The team at MIT say their system will be able to do these editors’ jobs in a faster, more efficient manner, all while maintaining language largely similar to how a human would write or edit.
At least for now, some humans are still necessary for the new MIT system to work properly. In execution, a human would type into an interface rough, accurate information without having to worry about forming full sentences or using proper grammar. Then, the system would search through Wikipedia, identify any inconsistencies, and rewrite the offending sentences to reflect accurate information written in a humanlike manner. In the future, the research team say they plan on improving the AI system to the point that no human input at all will be necessary.
“There are so many updates constantly needed to Wikipedia articles. It would be beneficial to automatically modify exact portions of the articles, with little to no human intervention,” says Darsh Shah, a PhD student in the Computer Science and Artificial Intelligence Laboratory (CSAIL) and one of the lead authors, in an MIT release. “Instead of hundreds of people working on modifying each Wikipedia article, then you’ll only need a few, because the model is helping or doing it automatically. That offers dramatic improvements in efficiency.”
While this isn’t the first bot or system developed to update Wikipedia articles, this is by far the most advanced. Earlier systems merely do away with blatant vandalism on Wikipedia or add very limited bits and pieces of information into already formed sentences. This new system, however, provides a much more complex service: taking a piece of unstructured information and instantly identifying where it is needed on Wikipedia and writing a new sentence.
“The other [bot] tasks are more rule-based, while this is a task requiring reasoning over contradictory parts in two sentences and generating a coherent piece of text,” Shah notes.
The system isn’t limited to Wikipedia either, researchers also used it to reduce bias in a popular fact-checking database. In fact, the study states that the system is useful in identifying fake news, an ever-present problem in today’s online landscape. More specifically, it can help train other AI systems designed to find and eliminate fake news by reducing bias.
Today, we all take websites like Wikipedia for granted, and it’s easy to forget that not that long ago an information source like Wikipedia would sound like something out of a science fiction novel. Now it seems that in a few years down the road, it very well may be.
The study is being presented at the AAAI Conference on Artificial Intelligence.