There is an obvious benefit to using face recognition for security, but many believe use by law enforcement and government agencies cross a fine line between helping and hurting.
There is an obvious benefit to using face recognition for security, but many believe use by law enforcement and government agencies cross a fine line between helping and hurting.
Face recognition has always has a bit of a stigma surrounding its use. While Hollywood makes it look cutting-edge and cool, in reality it feels a bit creepy that a person can be tracked with their face.
The idea of law enforcement agencies using face recognition technology to help find and identify criminals sounds like a good idea; however, inaccuracies mean innocent bystanders are getting IDed as suspects. Furthermore, many worry about the police abusing the technology.
In the U.S., news of Amazon’s face recognition, known as Rekognition, being used by US law enforcement sparked a conversation about responsible use of the technology. In response, Brian Brackeen, CEO of face recognition company
Karios, wrote a piece for the website TechCrunch, opposing the use of the technology by law enforcement agencies.
“In a social climate wracked with protests and angst around disproportionate prison populations and police misconduct, engaging software that is clearly not ready for civil use in law enforcement activities does not serve citizens, and will only lead to further unrest,” Brackeen wrote.
According to Big Brother Watch, in May of this year, 98 percent of the face recognition matches made by the Metropolitan Police (the Met) had wrongly identified innocent persons; Big Brother Watch is a British civil liberties organization that just recently issued a claim with the British High Court against the Met’s use of face recognition cameras. The group’s investigation claims that even when innocent people were wrongly identified the Met still stored biometric photos of the individuals without their knowledge for up to a year. As long as people continue to feel their rights are being violated, it is unlikely that those against law enforcement’s use of face recognition technology will ever support its use.
Addressing public concerns
While privacy is a concern for face recognition in general, companies say it’s not the main concern when it comes to law enforcement’s use of the technology. Roger Rodriguez, Retired NYPD Detective and Director of Client Relations at
Vigilant Solutions, stressed the importance of ensuring policy, accountability and transparency at the agency level, during and after deployments of face recognition systems.
“All parties (e.g., law enforcement, city leaders, legislators, privacy groups, etc.) must have a clear understanding of how this technology actually works and move away from the Hollywood glamorization of the technology,” Rodriguez said. “Only through education can everyone fully understand facial recognition’s current limitations.”
Furthermore, every deployment must have a self-regulating inquiry system which meets Criminal Justice Information Services (CJIS) compliance standards. According to Rodriguez, system wide audits and monitors for agency managers must cover the following: agency metrics, data sharing, gallery management, supervision of personnel, and alert generation and monitoring.
“Any agency electing to use facial recognition capabilities must document each search providing a historical audit of the process from start to finish which includes the reasons why the search is performed, enhancements made to the images and any facial annotations identified during the identification process,” Rodriguez explained.
Where the databases come from
One major concern comes from the databases being used by law enforcement to make matches. Enrollment in a criminal database is a highly sensitive action, and it should be connected to an actual crime.
“Law enforcement agencies should not be above the law, and feed such databases with facial images collected under a different consent. For example, ID photos collected for drivers’ licenses should not automatically transfer into a criminal database,” Elke Oberg, Marketing Manager at
Cognitec Systems, said. “In regards to biometric data and its use, police agencies should strive for responsible procedures addressing the issues of data security, data sensibility and rights to privacy. Law enforcement agencies should use transparent methods to inform the public about how biometric data is collected, stored, protected and eventually deleted.”
Rodriguez also stressed that any image enrolled into a database should always be documented as part of a bonafide law enforcement investigation. He explained that images enrolled into the facial recognition systems are sourced by investigators through video surveillance at the scene of the crime or are usually provided by a complainant or witness as part of an active investigation.