THE RIGHT TO EXPLANATION: A CONCEPT IN TENSION WITH COMPETING OBJECTIVES AND THE INTEGRITY OF AUTOMA
Data subjects (or “people” as commonly referred to in the technology community) are increasingly becoming subjected to more and more automated decisions that may evaluate a person’s worthiness for many services and opportunities. For instance, automated decisions resulting from algorithms are used in evaluating a person’s credit worthiness, performance at work, health, economic prospects, personal preferences and many other aspects of an individual.[1] Algorithms are used in rehabilitation and sentencing decisions related to a person’s propensity to commit future criminal acts, automated decisions are already well implemented in our society and the negative impact on many communities, especially underrepresented communities, is already underway.[2]
In fact, on February 16, 2017 the EU updated its current GDPR or General Data Protection Regulation. Seemingly, the EU seeks to include a Right to Explanation in an effort to minimize the potential for harmful discrimination, made without human intervention. As it stands now, many of the regulations in the United States and the EU are unclear or are silent on liability if harm to persons or property are done with little to no human intervention, as is the case with automated decisions and artificial intelligence.[3]
The EU has debated whether a Right to Explanation of an automated decision or artificial intelligence is necessary and has implemented safeguards within the regulation to prevent the potential for widespread harm. A Right to Explanation suggests that data subjects are entitled to an explanation of the automated decision, whether the decision is a result of the functionality of the system as a whole or an explanation of a specific decision as it affects a specific person and the factors and weight given to each factor included in determining the decision.[4]
Under the GDPR, the Right to Explanation includes that individuals have the right to not be the subject to a decision when the decision is based on automated processing and the decision produces legal effect or other significant effect on the individual. Still, the GDPR indicates that a Right to Explanation would not apply if explicit consent is obtained, if authorized by law, or if necessary for entering and performing a contract.[5]
As it relates to profiling, the GDPR suggests that when processing personal data for profiling purposes, appropriate safeguards must be in place to ensure that processing is fair and transparent by providing meaningful information about the logic involved, and any significant consequences. Further, the GDPR requires the use of appropriate mathematical or statistical procedures for the profiling, the implementation of appropriate technical and organizational measures to correct inaccuracies and minimize the risk of error, and requires that organizations processing personal data secure the data in a way that is proportional to the risk to interests and rights of the individual and prevents discriminatory effects.[6]
The GDPR also requires that automated decisions must not concern a child, or be based on the processing of special categories of data unless explicit consent is obtained or the processing is necessary for reasons of substantial public interest.[7]
As advances in algorithmic decision making and profiling continue, the public has a deepening interest in understanding how decisions that may negatively impact them are made. The ability to receive information about decisions that negatively impact an individual should not be obstructed, as accountability and transparency merely contributes to human and machine progress.
Read more here:
[7] Id.