Developing an AI that detects inappropriate builds

Discussion in 'Plugin Development' started by TurtlesBeast23, Jun 2, 2018.

Thread Status:
Not open for further replies.
  1. Offline

    TurtlesBeast23

    I was thinking of developing an AI that detects inappropriate builds which can include pixel arts or constructions of vulgar body parts or sexually-explicit and phallic builds, profanity and builds that may refer to or allude to an offensive symbol. The AI has features pre-programmed into it which is enough to automatically moderate servers, however, it also learns from other inappropriate builds that players create and smart ways they use to avoid detection.

    I'm wondering, would this be possible?

    Sincerely,
    TurtlesBeast23
     
    Last edited by a moderator: Jun 2, 2018
    TurtlesBeast23, Jun 2, 2018
    #1
  2. Offline

    MightyOne

    If you are advanced in AI programming there will surely be a way to teach one to detect inappropiate builds. Since you are asking I guess you have no certain knowledge about AI progrmming, right? Symbols and implications are not hard to recognise for our human brain. But already autistic people might have problems to interprete such. You might understand that it would take years to teach an AI interpretation of builds.
     
    Last edited: Jun 3, 2018
    MightyOne, Jun 3, 2018
    #2
  3. Offline

    timtower Administrator Administrator Moderator

    @TurtlesBeast23 @MightyOne Things to keep in mind as well:
    It needs to know what a specific build is, you can't just let it check the entire world, an area or exact building is the best way.
    It needs to be trained, this means gathering a training set of builds and a test set of builds.
    The more the better.
     
    timtower, Jun 3, 2018
    #3
Thread Status:
Not open for further replies.

Share This Page