Milestone 3 NotMathMajors DarkHorseIdea: Artificially Intelligent System for Submission Quality Assurance

From crowdresearch
Jump to: navigation, search

Initial Summary

Workers submissions are recorded by an intelligent machine during certain tasks to ensure authenticity and diligence. Work that fails the algorithm that determines authenticity would be flagged for review by requesters. The intelligent machine would learn from tasks that are completed and reviewed by the requester in order to adjust the algorithm for valid work. The "intelligent machine" could also be anonymous HITs that are paid for by a small tax from requesters.

Possible Workflow

AISSQANMM.png

  • Requestor creates task
  • Worker completes task
    • AI records worker input and time taken while the worker works.
  • Based on AI recording, AI either flags or passes the submission.
  • If flagged, the requestor will have the opportunity to reject the submission or accept it.
    • Based on this result, the AI will learn more about how to determine whether a submission is acceptable or not within that category.