Personal data protection and artificial intelligence in terms of transparency and accountability

Artificial intelligence systems are widely employed nowadays when making decisions that have impacts on individuals and society. The fact that AI systems learn from data sets does not guarantee that their outputs will be free of human bias or discrimination. With each new use case of AI, the risk of it causing inequality, discrimination and breaching data protection rights increases. It was necessary to investigate options about how algorithmic decision-making systems would engage in personal data processing that comply with data protection law. In this study, personal data protection and artificial intelligence systems in terms of transparency and accountability were investigated. As a research method, literature review method, which is one of the qualitative research methods, has been used since an existing situation will be identified. In this study, the types and usage areas of artificial intelligence systems, articles regarding profiling and automatic decision making mechanisms within the scope of GDPR have been included. Furthermore, the black box and opacity issues, the concepts of algorithmic bias and algorithmic discrimination and their relations with transparency and accountability have been evaluated. Finally, the Data protection impact assessments, the concept of data protection by design, data protection officer and emerging algorithmic audit mechanisms have been covered. Keywords: GDPR, General Data Protection Regulation, Accountability, Personal Data Protection, The Turkish Personal Data Protection Law, Profiling, Automated Decision-Making Mechanisms, Transparency, Artificial Intelligence.