It's a whole new world for customer experience coming. Get ready for the day when computers know you by sight. The FBI is.
The Federal Bureau of Investigation is rolling out its US$ 1 billion Next Generation Identification (NGI) system, whose technology includes facial recognition, voice recognition, iris and retina scans, and enhanced forms of old fashioned fingerprinting.
13 Million Mugshots
The system has capabilities that could thrill science fiction fans, alarm criminals and outrage privacy advocates.
With NGI, for example, the FBI can submit photos obtained from surveillance cameras to be matched with photos in its database, currently containing nearly 13 million mugshots. The Bureau is also developing a database of images of scars and tattoos for similar matchups. The agency has indicated that the facial recognition function of NGI successfully makes a match more than 90 percent of the time.
In July, the FBI’s Jerome Pender, Deputy Assistant Director of the Criminal Justice Information Services Division, told a subcommittee of the Senate Judiciary committee that a successful end-to-end Facial Recognition Pilot test has been conducted in Michigan.
Pender said that the database only contains photos of known criminals, and that images from such sources as social networking sites and general surveillance cameras were not used to populate the repository.
Privacy and Civil Liberties
His testimony was entitled “What Facial Recognition Means for Privacy and Civil Liberties,” an indication that the FBI recognizes, at least verbally, the potential for misuse. The Electronic Frontier Foundation, for instance, has asked Congress to limit the collection of biometric material and to require a warrant requirement of probable cause in order to use the system.
As of July, NGI was about 60% implemented. Full implementation is expected nationally by 2014.
NGI’s rollout actually began in March of 2010, with the deployment of advanced technology workstations, and has proceeded in stages through more accurate fingerprint searches, national deployment of mobile fingerprint IDs, facial recognition, palm searching and latent print searching, the scar and tattoo repository, and iris recognition.
Other fast-moving developments point to the day when facial recognition and other biometrics will not only be a common component of law enforcement, but of everyday life.
Figuring Out SSNs
Professor Alessandro Acquisti of Carnegie Mellon University also testified in July before the Senate. He noted the various uses of facial recognition that are popping up in products and services, such as face-matching between photos in Facebook, Apple’s iPhoto, and Google’s Picasa.
He also pointed to research conducted by his team, in which they were able to successfully predict the first five digits of Social Security Numbers for some individuals, in addition to other personal information, starting with an anonymous face and using data mining. “Within a few years,” he said, “real-time, automated, mass-scale facial recognition will be technologically feasible and economically efficient” for individual end-users.
And senior executives at Intel and Advanced Micro Devices have been making visionary presentations about the next phase of computer interaction. On Tuesday, Intel executive vice president Dadi Perlmutter described the oncoming wave of “perceptual computing,” built around facial recognition, voice recognition, eye-tracking and gestural controls. In August, AMD’s Chief Technology Officer, Mark Papermaster envisioned the coming of “surround computing,” similarly characterized by natural user interfaces based on voice and facial recognition.
Seems the customer experience will get just a little more personal.
- It's Official: Forrester Says Campaign Marketing Is Dead
- Adobe: IBM's Silverpop Deal Could Trigger 'Nightmare'
- Will EMC Dump Documentum?
- Hackers Use Viral Videos to Attack B2B E-Commerce Site
- A Beginner's Guide to Responsive Web Design
- Dream On Salesforce, SAP Prez Unimpressed by Your Threats
- Was IBM's Hand Forced by Industry to Acquire Silverpop?