A controversial UK murder prediction algorithm developed by government-linked researchers has come under fire from civil rights advocates, who warn it could lead to racial bias, profiling, and severe privacy violations.
According to documents uncovered by the group Statewatch, the algorithm has been in development for the past two years under a project initially called the “Homicide Prediction Project,” now renamed “Sharing Data to Improve Risk Assessment” (SDIRA).
The research involves a database of 100,000 to 500,000 criminal justice records, including data on convicted offenders, crime victims, and witnesses.
Dystopian Concerns Raised by Civil Rights Groups
The project is a collaboration between the UK Ministry of Justice, the Home Office, Greater Manchester Police, and the Metropolitan Police. Statewatch researcher Sofia Lyall called the program “chilling and dystopian,” noting that the algorithm uses sensitive personal data to predict the likelihood of individuals committing murder in the future.
The data includes not just basic identifiers like name and date of birth, but also ethnicity and “health markers” such as addiction, self-harm behavior, suicidal ideation, and other mental health issues. These details, officials say, are believed to have significant predictive power.
Ethical Concerns Over Mental Health Data Use
Records date back to before 2015 and were primarily gathered from Greater Manchester Police and the British Probation Service. The use of mental health and addiction-related data has been criticized as highly intrusive, with Lyall stating: “Building an automated tool to profile people as violent criminals is deeply wrong.”
According to Statewatch, such profiling carries the risk of reinforcing structural inequalities in the criminal justice system.
Fears of Bias Against Minorities and the Poor
Chris Jones, Director of Statewatch, argued the system could reproduce biases inherent in UK law enforcement data. “We already know that Black people, other ethnic minorities, and poor people are over-represented in police data,” he said. “Using that data to build predictive tools will only reinforce existing systemic racism.”
The Ministry of Justice confirmed the existence of the project but emphasized it was solely for “research purposes” aimed at understanding the risk of serious violent reoffending among people on probation. Officials insisted the data involved only individuals with a criminal conviction.
Disputed Claims and Lack of Consent
Jones dismissed the ministry’s claims as “wildly inaccurate,” citing documents indicating that data was pulled not just from offenders, but also from individuals labeled as victims, witnesses, missing persons, or those involved in safeguarding cases. “If that’s not profiling, I don’t know what is,” he said.
He also raised concerns that none of the individuals whose data was used had been informed, though officials stated the data had been anonymized and identification was not possible.
Despite being scheduled to end in December 2024, the controversial research project remains ongoing.
