Webinar - Criminal justice by algorithm: Predictive policing
This event is now finished. If you couldn't make it, or you want to watch again, you can watch the recording on our YouTube channel.
Speakers:
Host: Griff Ferris, Legal and Policy Officer, Fair Trials
Criminal justice and law enforcement decisions are increasingly being influenced and even made by machine-learning algorithms and automated-decision making systems, sometimes referred to as “artificial intelligence”.
One prevalent area is in policing and law enforcement, or ‘predictive policing’, which includes ‘predicting’ the likelihood of certain criminal acts occurring in a particular area, profiling individuals as likely to commit criminal acts in the future, or assessing people’s ‘risk’ of criminality. These systems are created and operated using criminal justice data, including police and crime records, but often other social data such as education, welfare or social security and other local authority records, and sometimes even financial or credit-broker records.
There are significant and fundamental issues with predictive policing systems and their use in criminal justice, not least because of the clear stereotyping and discrimination upon which these models are based, but also the unsurprisingly unjust and discriminatory decisions that these systems have been shown to generate. Meanwhile, law enforcement and government praise the ‘neutrality’ of such systems, leading to almost unchallengeable and structurally racist narratives of ‘criminal communities’.
These automated systems infringe fundamental rights: the right to a fair trial, especially the presumption of innocence, as well as privacy and data protection rights.
Speakers:
- Fieke Jansen, Data Justice Lab
- Dr Patrick Williams, co-author of ‘Data Driven Policing: the hardwiring of discriminatory policing practices cross Europe’ and ‘Being Matrixed: the (over)policing of gang suspects in London’.
- Sarah Chander, Senior Policy Adviser, European Digital Rights (EDRi)
Host: Griff Ferris, Legal and Policy Officer, Fair Trials
Criminal justice and law enforcement decisions are increasingly being influenced and even made by machine-learning algorithms and automated-decision making systems, sometimes referred to as “artificial intelligence”.
One prevalent area is in policing and law enforcement, or ‘predictive policing’, which includes ‘predicting’ the likelihood of certain criminal acts occurring in a particular area, profiling individuals as likely to commit criminal acts in the future, or assessing people’s ‘risk’ of criminality. These systems are created and operated using criminal justice data, including police and crime records, but often other social data such as education, welfare or social security and other local authority records, and sometimes even financial or credit-broker records.
There are significant and fundamental issues with predictive policing systems and their use in criminal justice, not least because of the clear stereotyping and discrimination upon which these models are based, but also the unsurprisingly unjust and discriminatory decisions that these systems have been shown to generate. Meanwhile, law enforcement and government praise the ‘neutrality’ of such systems, leading to almost unchallengeable and structurally racist narratives of ‘criminal communities’.
These automated systems infringe fundamental rights: the right to a fair trial, especially the presumption of innocence, as well as privacy and data protection rights.
* indicates required