Fed agency using Crowdflower & Amazon Mechanical Turk?

Home Forums Technology Fed agency using Crowdflower & Amazon Mechanical Turk?

This topic contains 5 replies, has 5 voices, and was last updated by  Steve Cottle 5 years, 2 months ago.

  • Author
  • #166470

    Justin Grimes

    Does anybody know of any examples of a federal agency using Amazon Mechanical Turk or Crowdflower?

  • #166480

    Steve Cottle

    Hi Justin,

    I haven’t heard of anyone using those specific services, but the State Dept’s Virtual Student Foreign Service was moving in the direction of building a cloud of student interns who could work on projects for embassies/missions around the world, like an in-house crowd-sourcing service. I wrote a short case in a larger report last year. You can find it on page 19 here.

  • #166478

    Steve Ressler

    Reach out to Crowdflower – I’d heard they had done some translation work during a couple crisis (I think Haiti) for either gov’t agencies or quasi-government groups

    I think while can make sense to do, obviously there’s a little sensitive of it for gov’t agencies based on people outside the U.S. doing the work (which I think you can control) and the wages (it’s almost too cheap/too good of a deal). But I think it’s an awesome way to go for certain tasks

  • #166476

    Chris Cairns

    This is a good question, Justin. I’ve been wondering lately ways that federal agencies could use a service like Crowdflower, Mechanical Turk and Microtask.

  • #166474

    John Hoskins

    Hi Justin;

    We have a couple good examples of Federal Agencies using and evaluating Mechanical Turk. Currently Army Research Labs used us to annotate short video clips for a DARPA project, DARPA is using us to adjust machine translation technology to recognize spoken language better. Finally, US Copyright Office is investigating digitizing and annotating its card catalog to bring those resources online.

    ARL used us directly, DARPA Language project is through Raytheon, and US Copyright is investigating several of our Partners for solutions.


  • #166472



    I’m Ari from CrowdFlower – a few examples for you:

    • NIH: identification and verification of drug and disease names in blocks of text, in order to complement their algorithmic NLP efforts
    • The US Copyright Office in transcribing registration cards, so they can be digitized and made public
    • Financial Crisis Inquiry Commission (FCIC): transcription, tagging and classification of documents (see more here: https://crowdflower.com/blog/2010/06/crowdsourcing-the-goldman-sachs-investigation/)

    Let me know if you’d like to learn more!


You must be logged in to reply to this topic.