Exploring algorithmic governance: The AI Act and new realities for criminal justice, and fundamental rights

The article attempts to investigate what remains of our comprehension of criminal justice and the protection of fundamental rights in an algorithmic society, highlighting the new challenges posed by artificial intelligence (AI) predictive systems. The integration of algorithms from the commercial and private sectors into the public sector, particularly in policing and law enforcement, is often justified by claims of enhanced efficiency and security. However, the present manuscript argues that utilising such tools from the public sector to assess and categorise individuals based on their likelihood of engaging in criminal activity or expressing antisocial behaviour undermines the fundamental principles of traditional criminal law. This can be understood since while prediction in other areas can be deemed valuable or even life-changing, in criminal justice, predicting the future on the basis of the past threatens to reify and reproduce existing inequalities of treatment by institutions. The analysis starts by examining the notion of algorithmic governance and justice, providing the necessary conceptual framework of how the domination of algorithms in public administration and governance can have a tremendous impact on the orientation of criminal justice and fundamental rights. A key objective of the article involves the critical examination of the AI Act from a theoretical perspective, in relation to the protection of fundamental rights and the promotion of democracy, but also the scrutiny of specific provisions related to the use of intrusive discriminatory AI systems, such as emotion recognition systems and predictive policing. Finally, the article emphasises that correct political decisions are necessary so that AI systems and techniques do not serve as tools of intrusive social control.