알고리즘 권력화와 규제 거버넌스

Abstract

Algorithm empowerment refers to platform operators tooling algorithms to increase their influence through algorithm control and control. The expression that algorithms become powerful is not correct because algorithms do not operate with consciousness on their own. Of course, controlling the algorithm does not help solve the problem because various problems arise by empowering the algorithm. Platform operators develop algorithms for corporate activities and use them to generate profits more efficiently. Of course, it is hard to deny that there are cases where algorithms are deliberately distorted in the process. As such, the need for regulation of algorithms is increasing in situations where various social problems arise as algorithms become power. Of course, even if there is a need for algorithm regulation, sufficient counterarguments are made and a process of persuasion is required accordingly. For example, operators view algorithm disclosure as a regulation and argue that trade secrets cannot be disclosed. However, can we think of releasing algorithms to the general public? There is no need for that. It is most important to resolve conflicts or misunderstandings in the policy or legislative process. Even if the purpose of the policy is justified, the legitimacy of the purpose may be undermined if procedural legitimacy is not secured. If bias and discrimination are reflected consciously or unconsciously in the process of planning, development, operation, and use of algorithms, constitutional values such as guaranteeing equality and basic rights can be undermined. To solve these limitations or problems, it is necessary to increase the ethics of the algorithm itself or strengthen the ethical or legal responsibility of humans involved. If discrimination is fixed by intentionally misusing or abusing algorithms, it is an act that undermines constitutional value and is subject to absolute regulation. Rather than presupposing the algorithm as a problem, the legislative plan for this is to come up with a way to block the wavelengths that the algorithm can bring, and it is a technology-neutral approach. Regulatory or legal approaches to securing fairness in algorithms are the last resort, and it is a policy decision to use collective intelligence or create a specialized institution for algorithms to find solutions to problems.