A first: biometrics used to sentence criminal

Published 1 February 2011

A judge ruled that biometric facial recognition could be submitted as evidence marking the first time such evidence has been used in a criminal trial; this move surprised many legal and scientific experts as facial recognition technology does not follow basic legal standards required for evidence; the decision may or not become a legal precedent as it was not made by a California appellate or supreme court

In early January, convicted murder Charles Heard received twenty-five years to life in a California prison for murder.

The case was unique because it was the first time that biometric facial recognition technology had been permitted to be used as evidence in the court room.

In the years following 9/11, DHS and cities around the world began experimenting with closed circuit security cameras and facial recognition software.

After spending billions of dollars into research and development, the results were disappointing.

According to SF Weekly, police departments in Florida, Virginia, and even Germany abandoned the use of such technology as it did not lead to any arrests while it was deployed.

In Germany officials found that the technology only had a 60 percent success rate in identifying people during the day, while at night it dropped to as low as 10 percent due to poor lighting conditions that made accurate identification difficult.

Given this track record, legal and scientific experts were surprised by a San Francisco Superior Court judge’s decision last July to allow biometric facial identification technology to be submitted as evidence to help exonerate a suspected murderer.

Surveillance cameras from a nearby business caught footage of man believed to have shot and killed another in an armed robbery.

The suspect’s defense team submitted still frames from the video footage along with testimony from a biometrics expert who argued that comparisons between current photos and the still frame clearly demonstrate that Charles Heard, the suspect, was not the shooter.

According to David Faigman, an expert on scientific evidence at Hastings College of Law, the judge’s decision came as a bit of a surprise because the technology does not meet a few basic legal standards as it applies to other forms of scientific evidence like DNA or fingerprint analysis.

I think it is precedent-setting,” he said, “But I also think that the appellate courts might take a dim view of the admission of this evidence…Without the systematic and rigorous evaluation of the evidence, it’s hard to know how much weight to give it.”

Faigman is troubled by the judge’s decision because of the unreliability of facial identification biometrics, especially if this becomes a precedent for future trials.

My experience suggests that it’s not generally accepted in the mainstream scientific community,” Faigman said.

The lead prosecutor, Assistant District Attorney Michael Swart, agreed with Faigman’s assessment as he argued against permitting such evidence in the court room.

In a motion, Swart said “At this point, it cannot be established that biometrics has been generally accepted in the relevant scientific community.”

He added, “Just as importantly, there are no prior published Court of Appeal opinions in California that establish the general scientific acceptance of this science.”

Heard was eventually convicted by a jury despite the biometric evidence showing that he could not have been the shooter.

The case’s precedent is not legally binding because it was not ruled on by California’s appellate or supreme court, leaving the future of facial-identification technology in court rooms uncertain.