In recent years, there is a growing use of algorithmic law enforcement by online intermediaries. Algorithmic enforcement by private intermediaries is located at the interface between public law and private ordering. It often reflects risk management and commercial interests of online intermediaries, effectively converging law enforcement and adjudication powers, at the hands of a small number of mega platforms. At the same time, algorithmic governance also plays a critical role in shaping access to online content and facilitating public discourse. Yet, online intermediaries are hardly held accountable for algorithmic enforcement, even though they may reach erroneous decisions. Developing proper accountability mechanisms is hence vital to create a check on algorithmic enforcement. Accordingly, relying on lessons drawn from algorithmic copyright enforcement by online intermediaries, this chapter demonstrates the accountability deficiencies in algorithmic copyright enforcement; maps the barriers for algorithmic accountability and discusses various strategies for enhancing accountability in algorithmic governance.
Verifying Controllers Against Adversarial Examples with Bayesian Optimization
