Iventory your algorithmic systems

Suggested time: 5 minutes

The first step in implementing algorithmic transparency is to understand everywhere that algorithmic decision-making systems are being used within your organization.


Create a list of all the algorithmic decision-making systems in your organization

The definition of an algorithmic system is quite broad, and includes any system that relies on the analysis of data to derive algorithms for making decisions. Other names for these tools are automated decision systems, artificial intelligence (AI) systems, machine learning (ML) systems, or data science systems. In general, algorithmic tools receive a data input, process that input through an algorithm, and then produce an output in the form of a decision.

From here forward, we will refer to algorithmic systems in shorthand as either algorithmic decision systems, algorithmic systems, or just simply algorithms. In all of these cases, we are referring to the decision-making system as a whole.


Examples of algorithmic systems include hiring tools, predictive algorithms, and smart metering systems. Predictive algorithms have a wide range of applications including predicting student performance, chance of hospital re-admission, risk of diseases, whether not a loan applicant should be accepted or rejected, or the likelihood a user will click on an advertisement.

When inventorying your organization for algorithms, keep these two important points in mind:

  • Algorithms are not always big, complex systems. Sometimes algorithmic systems can be deceptively simple — and its important not to overlook them! For example, many schools use Early Warning Systems that only look at three factors (a students’ grades, attendance record, and the number of detentions they recieved) which are tallied together, sometimes on paper and without the help of a computer. Despite its simplicity, this is still an algorithm!

  • Algorithms in procured tools. Many organizations, especially in governmental or intergovernmental organizations, choose to procure their algorithmic tools instead of building them in-house. Don’t forget to include these in your inventory! Even if they are not internal to your organization, there are still actions towards transparency for these type of tools.

DELIVERABLE: An inventory list of all the algorithms being used by your organization, including those that are procured.