Algorithmic Transparency and Accountability: Legal Approaches to Solving the "Black Box" Problem

The paper examines the European and American legal approaches based on legislation regulating the use of computer algorithms, i.e. systems for automated decision-making of legally significant decisions. It is established that these jurisdictions apply essentially different concepts.
The European approach provides for regulating the use of automated decision-making systems through legislation on personal data. The authors conclude that the general data protection regulation does not impose a legal obligation on the controllers to disclose technical information, i.e. to open a "black box", to the subject of personal data, in respect of which the algorithm makes a decision. This may happen in the future, when the legislative authorities specify the provisions of this Regulation, according to which the controller must provide the subject of personal data with meaningful information about the logic of decisions taken in relation to it.

In the United States, issues of transparency and accountability of algorithms are regulated by various antidiscrimination acts that regulate certain areas of human activity. At the same time, they are fragmentary and their totality does not represent a complex, interconnected system of regulatory legal acts. In practice, legal regulation is carried out ad hoc with reference to certain legal provisions prohibiting the processing of sensitive types of personal data.
The paper states that the legal regulation of algorithmic transparency and accountability is in its infancy in Russia. The existing legislation on personal data suggests that the domestic approach to solving the "black box" problem is close to the European one. When developing and adopting relevant regulatory legal acts, it is necessary to proceed from the fact that the subject of personal data should have the right to receive information explaining the logic of the decision made in relation to itin an accessible form.

1. Bruckner MA. Promise and Perils of Algorithmic Lenders’ Use of Big Data. Chicago-Kent Law Abstract. 2019;93(1). (In Eng.)

2. Burrell J. How the machine «thinks». Understanding opacity in machine learning algorithms. Big Data & Society. 2016;1-12. (In Eng.)

3. Isak M, Lee AB. The Right Not to Be Subject to Automated Decisions Based on Profiling. In: Synodinou T, Jougleux P, Markou C, Prastitou T, editors. EU Internet Law: Regulation and Enforcement. Springer; 2017. (In Eng.)

4. Malgieri G. Automated decision-making in the EU Member States: The right to explanation and other «suitable safeguards» in the national legislations. Computer law & security review. 2019;35. (In Eng.)

5. Malgieri G, Comandé G. Why a Right to Legibility of Automated Decision-Making Exists in the General Data Protection Regulation. International Data Privacy Law. 2017;7(4). (In Eng.)

6. Selbst AD, Powles J. Meaningful information and the right to explanation. International Data Privacy Law. 2017;7(4). (In Eng.)

7. Talia BG, Jann S. Big Data and Discrimination. University of Chicago Law Abstract. 2019;459. (In Eng.)

8. Wachter S, Mittelstadt B, Luciano Floridi L. Why a Right to Explanation of Automated Decision-Making Does Not Exist in the General Data Protection Regulation. International Data Privacy Law. 2017;7(2). (In Eng.)