UK government tests AI to predict murders

The UK government is developing an AI program designed to predict which individuals are most likely to commit serious violent crimes. As reported by The Guardian and revealed through documents obtained by the watchdog group Statewatch, the initiative aims to increase public safety by analysing data from individuals already known to authorities.

The system uses personal data, including information about mental health, addiction, and suicidal tendencies, to identify so-called “potential murderers.” The project has been widely criticised by civil rights advocates. Statewatch researcher Sofia Lyall warns that such systems risk reinforcing structural biases embedded in the criminal justice system.

While predictive policing is not new, its use remains highly controversial. Statewatch describes the initiative as “chilling and dystopian.” Similar concerns exist elsewhere. In the Netherlands, the national Court of Audit criticised the Criminal Anticipation System (CAS) in 2022 for potentially embedding biases against certain groups. CAS is still in use.

The UK Ministry of Justice states the project is still in a research phase and that only the data of convicted individuals is being used, particularly to assess reoffending risk during probation. However, the project is under scrutiny—especially as the European AI Act classifies predictive policing as an “unacceptable risk.” As such, this type of AI use will be banned within EU member states.

Read the full article on the website of de Volkskrant.


(Yes, this photo was generated with AI).