New Statement on Algorithmic Transparency and Accountability by ACM U.S. Public Policy Council

By Renee Dopplick, ACM Director of Public Policy
January 14, 2017

Recognizing the ubiquity of algorithms in our daily lives, as well as their far-reaching impact, the ACM US Public Policy Council issued a “Statement on Algorithmic Transparency and Accountability,” containing a list of seven principles designed to address potential harmful bias.

Algorithms, the set of instructions computers employ to carry out a task, influence almost every aspect of society. The explosive growth of data collection, coupled with increasingly sophisticated algorithms, has resulted in a significant increase in automated decision-making, as well as a greater reliance on algorithms in human decision-making. Industry forecasters believe software programs incorporating automated decision-making will only increase in the coming years as artificial intelligence becomes more mainstream. One of the major challenges of this emerging reality is to ensure that algorithms do not reinforce harmful and/or unfair biases.

In the related press release, ACM U.S. Public Policy Council Chair Stuart Shapiro said, “Algorithmic bias can occur even with the best of intentions.” He explained, “This is, in part, due to the fact that both software development and its products can be complex and produce unanticipated results. Following these principles cannot guarantee that there will be no biased algorithms or biased outputs. But they will serve to keep computing professionals on the lookout for ways biases could creep into systems and provide guidelines on how to minimize the potential for harm.”

The Statement on Algorithmic Transparency and Accountability was designed to be consistent with the ACM Code of Ethics.

The effort was initiated by the Algorithmic Accountability Working Group, which is Co-Chaired by Simson Garfinkel, Jeanna Matthews, and Jonathan M. Smith. Andy Oram assisted the Working Group Co-Chairs with the planning of a technical workshop that helped inform the statement. Contributors and reviewers included a high-level experts group of computing professionals, scientists, researchers, educators, and other technology professionals with backgrounds in a range of computing disciplines.

7 Principles for Algorithmic Transparency and Accountability

  1. Awareness

    Owners, designers, builders, users, and other stakeholders of analytic systems should be aware of the possible biases involved in their design, implementation, and use and the potential harm that biases can cause to individuals and society.

  2. Access and Redress

    Regulators should encourage the adoption of mechanisms that enable questioning and redress for individuals and groups that are adversely affected by algorithmically informed decisions.

  3. Accountability

    Institutions should be held responsible for decisions made by the algorithms that they use, even if it is not feasible to explain in detail how the algorithms produce their results.

  4. Explanation

    Systems and institutions that use algorithmic decision-making are encouraged to produce explanations regarding both the procedures followed by the algorithm and the specific decisions that are made. This is particularly important in public policy contexts.

  5. Data Provenance

    A description of the way in which the training data was collected should be maintained by the builders of the algorithms, accompanied by an exploration of the potential biases induced by the human or algorithmic data-gathering process. Public scrutiny of the data provides maximum opportunity for corrections. However, concerns over privacy, protecting trade secrets, or revelation of analytics that might allow malicious actors to game the system can justify restricting access to qualified and authorized individuals.

  6. Auditability

    Models, algorithms, data, and decisions should be recorded so that they can be audited in cases where harm is suspected.

  7. Validation and Testing

    Institutions should use rigorous methods to validate their models and document those methods and results. In particular, they should routinely perform tests to assess and determine whether the model generates discriminatory harm. Institutions are encouraged to make the results of such tests public.

ACM U.S. Public Policy Council

The ACM U.S. Public Policy Council is chartered as the focal point for ACM’s interaction with the U.S. government, the computing community, and the public in all matters of U.S. public policy related to computing and technology. USACM represents a diverse community of computing practitioners, scientists, educators, researchers, and other technology professionals from government, business, academia, and the nonprofit sector. Its contributions to public policy draws from the deep scientific and technical expertise of the computing community.