Are we sleepwalking into an AI police state?
Predictive analytics enabling law enforcement to identify “high-risk” areas has highlighted ethical and legal quandaries
Science fiction has long speculated on the danger of a dystopian future and machines powered by artificial intelligence (AI). But with the advent of big data, we no longer need to speculate: the future has arrived. By the end of March, West Midlands Police is due to finish a prototype for the National Data Analytics Solution (NDAS), an AI system designed to predict the risk of where crime will be committed and by whom. NDAS could eventually be rolled out by every police force in the UK.
Ultimately, we need to be able to choose as a society how we use these technologies and what kind of society we really want to be
Fourteen police forces around the UK have used or planned to use such tools. But a report published in February by human rights group Liberty warns that far from being objective, police crime-mapping software reinforces pre-existing biases about who commits crime.
Current mapping tools use past crime data to identify so-called high-risk areas, leading to more intensive patrolling. Yet these areas are often already subject to disproportionate over-policing. By relying on data from police practices, according to Liberty’s advocacy director Corey Stoughton, these tools might simply “entrench discrimination against black and minority-ethnic people”.
Police mapping tools turning citizens into suspects
Ben Hayes, a data protection and ethics adviser to the European Union, United Nations and other international organisations, warns that the increased use of such mapping tools is increasingly turning ordinary citizens into suspects.
“People can be categorised as vulnerable, at risk, threatening, deserving or undeserving,” says Dr Hayes, noting that this tends to target those already marginalised. “Services such as border control, policing and social welfare are all subject to inherent bias. Machine-learning doesn’t eliminate those biases, it amplifies and reinforces them.”
…click on the above link to read the rest of the article…
