The product, from California-based Veritone, represents not only the newest way for law enforcement to employ artificial intelligence but could help provide a safeguard against worries about privacy invasion, according to a company executive.
The new offering will come through the company’s Veritone Track, Jon Gacek, general manager of Veritone Public Safety, told Government Technology.
That application uses AI to identify people as “human-like objects,” he said. That means it avoids the use of facial recognition and other biometric markers and instead relies on such traits as, say, the color of a person’s shoes and clothes, or whether they are wearing sunglasses.
Such features can be mentioned in alerts, used in photo arrays and otherwise help investigators find suspects or missing persons.
One of the ideas driving that approach is to avoid any backlash against facial recognition — some cities have banned the tech’s use for police, though it remains unclear how well those bans are working — and to avoid as much as possible any worries about the potential for AI to invade personal privacy.
“We are using AI in a non-facial recognition way, to bring together massive amounts of evidence,” Gacek said.
The next step of that effort involves vehicles, he said.
As Gacek explained it, the tool will enable police to conduct more detailed searches for automobiles. Instead of, say, looking for a white F-150 pickup truck via make, model and year, agencies can use Veritone Track to seek a specific F-150 with 22-inch rims, tinted windows and bumper stickers.
The expansion of Veritone Track follows another move to bring more AI into law enforcement by the company.
It recently said that a “primary state law enforcement agency” will use Veritone Redact to automate via AI audio and video redactions — part of the Freedom of Information Act process. The technology can reduce the manual labor that goes into redaction, speeding up the time it takes for information to be released to the public.
That reflects a larger trend in the industry, with technology suppliers helping agencies reduce FOIA backlogs.
