Over a million developers have joined DZone.
{{announcement.body}}
{{announcement.title}}

AI and Criminal Interrogations

DZone's Guide to

AI and Criminal Interrogations

Will AI support criminal interrogations in the future?

· AI Zone ·
Free Resource

Bias comes in a variety of forms, all of them potentially damaging to the efficacy of your ML algorithm. Read how Alegion's Chief Data Scientist discusses the source of most headlines about AI failures here.

Law enforcement officers have the right to question anyone for any reason, but the person being questioned always has the right to leave. An interrogation is conducted when a person is suspected to have committed a specific crime. Unlike being questioned by police, during an interrogation, the suspect isn’t free to leave.

Criminal interrogations are conducted to produce a confession. The U.S. Supreme Court allows police to use deception to obtain a confession, but threats of physical intimidation make a confession inadmissible. Police know how to break down a suspect to get a confession, but it can take anywhere from 6 to 24 hours straight.

Unfortunately, the desire to get a confession can override an officer’s sense of integrity, causing them to push so hard they end up with a false confession. If police were able to use AI software to analyze a suspect’s micro-expressions and body language in real time, they might be able to get authentic confessions faster and be warned when they’re pushing an innocent person over the edge.

AI Would Bring Hope to the Falsely Accused

False confessions lead to false convictions and innocent people serving time. In the United States, as of 2014, 3,944 years had been served in prison by innocent people. Since 1989, more than 300 convictions have been overturned due to DNA testing. According to the Center for Wrongful Convictions, 52 percent of exonerations involved false imprisonment or perjury, and 42 percent involved official misconduct. These stats prove that the justice system isn’t perfect.

Implementing AI software to monitor interrogations would help but would require filming interrogations first. Currently, interrogations are only required to be filmed in 24 states.

Florida is currently debating making it state law to film interrogations. If passed, the new law would support those with limited mental capacity who often falsely admit guilt while under duress. In 2009, for example, a Florida man named Anthony Caravella was released from prison after serving 26 years for a murder he didn’t commit, Intelligent Video Solutions explains. He was 15 when he was convicted and had an IQ of 67. He was later awarded $7 million after a jury discovered he was framed. If Caravella’s interrogation had been filmed and processed through AI software, someone could have saved him from going to prison in the first place.

Caravella’s case prompted the Florida Supreme Court to create the Florida Innocence Commission to address this issue. In 2012, the FIC created standards for law enforcement officers to follow while interrogating suspects to reduce the number of wrongful convictions. One of those standards was to record statements made by suspects. The standard has not yet been enacted as law.

Police Are Moving in the Right Direction With Predictive Analysis

Police are already analyzing bodycam footage to predict criminal behavior. Taser, a company that provides bodycams and data storage to law enforcement, is building an in-house “AI team.” Taser stores terabytes of police footage on Evidence.com on private servers accessible only to police agencies for a monthly fee. Until now, the footage hasn’t been used for investigative purposes. Taser is collaborating with other companies to create software that will detect emotions and interpret body language from bodycam footage without human intervention. They’re also looking to use this footage to anticipate what may happen in the future.

The deep Machine Learning that powers AI software recognizes patterns and infers associations from large batches of data. The more data that gets collected, the more accurate those inferences become.

AI Can Help, but It’s up to Police Agencies to Implement It

The AI technology needed to assist with interrogations already exists. Tech companies have launched AI software that reads body language in real-time to determine what mood a person is in, with plans to use it for behavioral diagnosis. Implementing this type of software during an interrogation would provide immense benefit to police, and the fact that it would be filmed would keep them honest about their tactics.

There’s one big roadblock to using AI to analyze interrogations: Many law enforcement officers believe filming an interrogation makes it harder to obtain a confession. If they don’t want to film interrogations, they can’t use AI software.

The absence of video surveillance during an interrogation is an unfortunate opportunity for police to use illegal tactics to obtain a confession — illegal tactics that put thousands of innocent people behind bars. However, with AI, police wouldn’t have to work so hard. Perhaps the benefits of using artificial intelligence will convince more law enforcement agencies to film interrogations and, hopefully, keep innocent people out of jail.

Your machine learning project needs enormous amounts of training data to get to a production-ready confidence level. Get a checklist approach to assembling the combination of technology, workforce and project management skills you’ll need to prepare your own training data.

Topics:
criminal interrogations ,artificial intelligence ,ai and crime ,predictive analysis

Published at DZone with permission of

Opinions expressed by DZone contributors are their own.

{{ parent.title || parent.header.title}}

{{ parent.tldr }}

{{ parent.urlSource.name }}