ASA Connect

 View Only

  • 1.  NIST draft Proposal for Identifying and Managing Bias in Artificial Intelligence

    Posted 07-08-2021 11:28

    Dear ASA Community, 

    The National Institute of Standards and Technology (NIST) has release a draft Proposal for Identifying and Managing Bias in Artificial Intelligence: https://nvlpubs.nist.gov/nistpubs/SpecialPublications/NIST.SP.1270-draft.pdf. They encourage organizations to review it and provide feedback by August 5. If you have recommendations, let me know. 

    Thanks to Kelly Zou for bringing this to my attention. 

    Steve




    ------------------------------
    Steve Pierson
    Director of Science Policy
    American Statistical Association
    ------------------------------


  • 2.  RE: NIST draft Proposal for Identifying and Managing Bias in Artificial Intelligence

    Posted 03-16-2022 08:32
    To close the loop on this, NIST has finalized its publication, per email announcement yesterday (below). They also announce a workshop on AI at the end of the month. 
    Steve

    Colleagues

    I am happy to share the just published NIST SP 1270 "Towards a Standard for Identifying and Managing Bias in Artificial Intelligence" revised on public comments. The updated document takes a socio-technical approach to AI bias. Please check it out in its entirety and consider joining us for the NIST AI workshop March 29th-31st, in particular the third day which is dedicated to discussions of bias in AI.

     

    A few of the authors' favorite takeaways:

    -There is more to bias than statistical and computational factors

    -Look at the data, carefully

    -Bias is treated most effectively within a given operational context

    -Follow the scientific method in AI activities

    -Apply human centered design to AI systems

    -Form governance structures for the humans that build and oversee AI systems

     

    Happy reading and please consider sharing this widely with your AI colleagues.

    Thank you

     

    NIST AI Bias page https://www.nist.gov/artificial-intelligence/ai-fundamental-research-free-bias

    Email us at ai-bias@list.nist.gov



    ------------------------------
    Steve Pierson
    Director of Science Policy
    American Statistical Association
    ------------------------------

    Original Message


  • 3.  RE: NIST draft Proposal for Identifying and Managing Bias in Artificial Intelligence

    Posted 03-25-2022 08:50
    A related NIST document, an initial draft of the AI Risk Management Framework, is now available for comment through April 29, 2022.
    https://www.nist.gov/itl/ai-risk-management-framework

    A two-part NIST workshop from March 29-31, 2022, will advance work on an AI Risk Management Framework – and on bias in AI. Register Now.

    NIST is developing a framework to better manage risks to individuals, organizations, and society associated with artificial intelligence (AI). The NIST Artificial Intelligence Risk Management Framework (AI RMF or Framework) is intended for voluntary use and to improve the ability to incorporate trustworthiness considerations into the design, development, use, and evaluation of AI products, services, and systems.

    The Framework is being developed through a consensus-driven, open, transparent, and collaborative process that will include workshops and other opportunities to provide input. It is intended to build on, align with, and support AI risk management efforts by others. An initial draft of the AI RMF is available for comment through April 29, 2022.

    NIST's work on the Framework is consistent with its broader AI efforts, recommendations by the National Security Commission on Artificial Intelligence, and the Plan for Federal Engagement in AI Standards and Related Tools. Congress has directed NIST to collaborate with the private and public sectors to develop the AI RMF.

    The Framework aims to foster the development of innovative approaches to address characteristics of trustworthiness including accuracy, explainability and interpretability, reliability, privacy, robustness, safety, security (resilience), and mitigation of unintended and/or harmful bias, as well as of harmful uses. The Framework should consider and encompass principles such as transparency, accountability, and fairness during pre-design, design and development, deployment, use, and test and evaluation of AI technologies and systems. These characteristics and principles are generally considered as contributing to the trustworthiness of AI technologies and systems, products, and services



    ------------------------------
    Steve Pierson
    Director of Science Policy
    American Statistical Association
    ------------------------------

    Original Message


Related Content