Editorials

Pull the plug

The short history of facial recognition technology in Ohio does not warrant a premature green light

8/30/2013
DeWine
DeWine

There's little doubt that Ohio’s digital facial recognition program will help law enforcement officers identify suspects and solve crimes. But like all other law enforcement tools, however valuable, it also carries a potential for abuse.

State Attorney General Mike DeWine has, prudently, appointed an advisory group to recommend guidelines and rules to prevent misuse, keep information secure, and protect privacy. He’s asking not only for recommendations on the use of facial recognition software, but also broader guidelines for operating the Ohio Law Enforcement Gateway. OHLEG is the computerized network used by law enforcement in the state to share data, including criminal histories, evidence, and missing-child information.

Click here to read more Blade editorials.

The advisory group is expected to make its recommendations in less than 60 days. Mr. DeWine said in an interview with The Blade’s editorial page that he’s committed to carrying out all of them.

Until then, Mr. DeWine should turn off the new facial recognition software.

The public has valid concerns about privacy and abuse. Those concerns have been aggravated by the failure of Mr. DeWine’s office to notify Ohioans when the system was activated on June 6, and by the broader controversy over the reach of National Security Agency databases.

Since June, police around the state have used the facial recognition system almost 2,700 times. Nearly 30 other states and the FBI use similar technology.

The new program enables law enforcement digitally to compare Bureau of Motor Vehicle photos to images captured by security cameras and other sources, speeding up the process of identifying suspects. Previously, law enforcement officers relied on names and addresses to search the state database of license photos and mug shots.

Mr. DeWine makes good arguments for the system, and Ohioans should expect him to turn the system back on — after the task force has issued rules and guidelines for operating the program.

As Mr. DeWine points out, facial recognition technology does not bring new evidence or ID photos to the BMV database, which law enforcement has used for 30 years. But it does enable police to use BMV identification photos more efficiently, by electronically comparing them with images of people taken off security cameras and other devices.

Facial recognition software digitally compares key features of a person’s face — sometimes called “points” — with driver’s license photos, looking for possible matches. The straight-ahead facial images captured on driver’s licenses are easy to compare with images of suspects on security cameras and cell phones.

Facial recognition technology can provides leads for police to follow, but it does not provide evidence as exact as DNA or fingerprints. There are margins of error and, therefore, abuse.

Without limits on what government can do with facial recognition software, containing this technology to criminal suspects would be practically impossible. By law, only law enforcement personnel can use OHLEG — and then only for police purposes. Still, numerous prosecutions for violating that law have occurred over the past decade.

Mr. DeWine concedes he erred in not announcing the facial recognition program, reported this week by the Cincinnati Enquirer.

“People are much more sensitive to all of this today,” he told The Blade editorial page. “I fully get it.”

The attorney general said he believes existing safeguards are adequate for the use of facial recognition software. Yet he has properly appointed a task force charged with improving them. By doing so, he has implicitly acknowledged potential problems, or at least better ways of managing them.

“The key is to build in enough safeguards,” he told The Blade. “If there is a violation, there have to be penalties strict enough to serve as a deterrent.”

Therein lies the problem: Ohio does not yet know whether existing safeguards are sufficient for the new technology, whose power is compounded exponentially by society’s ubiquitous security cameras. The few examples Mr. DeWine has cited to show the system’s effectiveness are not compelling enough to risk a premature green light.

Technology often brings new promises and perils. Facial recognition software is no different. It’s worth waiting 60 days for Ohio to get it right.