Criminal justice algorithms can now inform judicial decisions by foretelling future criminal behaviour. More concretely, they can contribute to the evaluation of the accused by assessing the risk of reoffending. In the name of correct risk foreseeing, these technologies may use any supposedly accuracy-enhancing factor. Criteria, including financial status, gender or age, may be considered as valid for the purpose of calculating probabilities; albeit, they may be neither blameworthy per se nor controllable by the defendant. Such algorithmic implementations have raised serious concerns. Different individuals accused of having committed the same criminal offence may be treated/punished in a different way. Furthermore, the defence may be unaware of the application of such risk assessment tools that can moreover be unchallengeable, due to their proprietary nature and/or unintelligible decision-making.

• How can traditional approaches to crime and/or punishment justify the application of these risk assessment tools?
• What is the state-of-the-art of these technologies in terms of regulation, practice and performance and how prepared is the EU to introduce these tools into criminal courts?
• To what extent can the right to the protection of personal data be interfered with by these technologies and how could the GDPR and the Data Protection Directive for Police and Criminal Justice Authorities safeguard this right?
• Which defence rights are at stake and how effective, but also desirable, could the possible contribution of data protection laws to the regulation of criminal areas be?