Regulating Artificial Intelligence To Protect The Right To Be Forgotten

Abstract

This dissertation attempts to answer the research question: can Artificial Intelligence (AI) be regulated to protect the Right to be Forgotten (RTBF)? It argues that although regulating AI to protect the RTBF is associated with several challenges, it can be regulated. While there is no concrete definition of AI, it can be understood through other means (classifications and differentiation). Based on the conception of AI, implement the RTBF in AI is very complex, because AI mainly consists of connected big data, deleting some of these data could cause inaccurate and undesirable results against the AI purpose. The RTBF has been developed over the years by EU laws, court and finally the GDPR, which explicitly articulates for the RTBF and increases its protection. Besides these developments, the territorial scope of protecting the RTBF is not solved. Regulating AI to specifically protect the RTBF is technically challenging, due to the vagueness of how AI ‘forgets’, the ease of requesting data deletion, the difficulty of the deletion process and the need for human intervention to evaluate the legitimacy of the request. Balancing between the RTBF and other public interests is also difficult. Based on the law, courts and practice, balancing between the RTBF and other public interest is subjective and thus automation may be futile in its current state. Nevertheless, AI could be regulated by learning from cyberspace lessons such as using code in Lessig’s modalities, which refers to the physical constraints that remove the individual’s ability to choose whether to comply or not. Although code can perhaps work better in regulating AI in general, regulating AI using algorithmic regulation depending on machine learning would be more efficient to protect the RTBF. Moreover, regulatory strategies can be utilised to effectively regulate AI to protect the RTBF, especially in areas where code is insufficient.