It was something necessary, has been too slow in coming. This is what I thought when I seen how in his latest press release released by the Google marketing department announces that the company is currently developing a new artificial intelligence system capable of moderating all comments “toxic” present in any forum or community.
This new project has been baptized under the name of Perspective and is nothing more than a new platform, I insist that still in phase of development, able to use machine learning to process all kinds of comments automatically giving them a score of 0 to 100 depending on whether they are truly building or the opposite. Undoubtedly a tool that surely will help, the less, limiting that power can have a user when it is expressed after a computer screen.
Perspective, the new platform of artificial intelligence of Google to moderate offensive comments.
Now, talk to you about that Perspective is just a support artificial intelligence system, i.e., once the platform has valued and rated a comment, remains in the hands of every editor decide what to do with it. As detail, let know you that this decision can also be automated in such a way that you can only publish messages with higher scores on each platform, messages with scores lowest, launch the author a notice indicating that your comment may be offensive… this decision will have to take it every administrator.
Without a doubt a very interesting tool, especially for communities too large volume of comments makes them impossible to moderate by any person, in the words of the own Google:
Perspective analyzes the reviews and evaluates them based on the resemblance that have comments that people have indicated as “toxic” or, probably, they can do that someone leave a conversation. To learn how to detect a potentially toxic language, Perspective examines tens of thousands of comments tagged by people dedicated to review them. Whenever Perspective find new examples of toxic comments or receives corrections by users, improves their ability to assess the future comments.