Palantir's secretive predictive policing programme in New Orleans


Under a secret deal beginning in 2012, the data mining company Palantir provided software to a New Orleans Police Department programme that used a variety of data such as ties to gang members, criminal histories, and social media to predict the likelihood that individuals would commit acts of violence or become victims. This partnership was extended three times through February 21, 2018. Even city council members had no idea the arrangement existed even though Palantir has used its New Orleans work as a reference for a multi-million dollar contract with another agency. Political commentator James Carville, a paid advisor to Palantir, says he is the driver of this project, which Palantir describes as philanthropy. Critics believe the programme was kept secret because American Southerners value their privacy, and it would have inspired widespread outrage.

From company documents, external studies, and the recollections of former NOPD chief Ronal Serpas, it appears that Palantir's prediction model in New Orleans used social network analysis to find connections between people, places, cars, weapons, addresses, social media posts, and other indicators that were held in previously separated databases, as well as field interview cards. NOPD used the list of about 3,900 potential victims and perpetrators to target individuals for the CeaseFire programme, which offers potential offenders of criminal records various services, but also threatens them with the fullest possible prosecution should they reoffend. Although there was a temporary drop in violent crime it's not clear that Palantir's service was responsible. The secrecy meant that attorneys representing defendants may not have been provided with evidence they had a right to see. 

In 2016, Palantir signed a seven-year contract to provide predictive technology aimed at identifying potential terrorists with the Danish national police and intelligence services.
Writer: Ali Winston
Publication: The Verge

Related learning resources
Target Profile