intelligence-powered body cams are almost ready for the men in blue,
which will provide an essential upgrade for fighting crime in the 21st
Axon recently announced
it would be embedding artificial intelligence in its wearable body
cameras. The company is known for the wearable body cameras and Taser
electroshock weapons used by most police departments in the U.S.
These body cams will have AI face recognition technology in them,
which will allow the police to scan and recognize the faces of almost
everyone they see while on patrol.
The technology is still in its early stages. The new algorithm
embedded in the cameras is designed to spot suspects and missing
children in real time. Once spotted, it notifies the officer who will
then decide on his next course of action. This could save hundreds of
man-hours in search and follow procedures. For missing children and
teens, this could save lives.
However, privacy advocates
are crying foul, as the technology could be a pervasive violation of
privacy. They feel that in the wrong hands, this kind of surveillance
could endanger innocent individuals.
Axon is the not only private tech company focused on AI for wearables.
Motorola Solutions has also recently partnered with Neurala, an AI
Motorola is an established vendor for communications tools and body camera technology for law enforcement agencies. With Neurala’s AI technology, Motorola will now create body cams that can help locate missing persons or identify suspects.
The trend is rapidly evolving, even outside of law enforcement. A
growing number of tech start-ups and surveillance firms are integrating
AI and other face recognition capabilities in real-time video.
How does it work?
A central database is loaded with all kinds of information and images
of a suspect or a missing person. Demographic details like gender, age,
hair color will be a part of the record. Characteristics like lisps,
moles and other distinctive features will be filled in as well.
The AI parses information from this database and stores it in the body
cam. The software is expected to get smarter over time by collating and
absorbing more data.
The "deep-learning" algorithms analyze facial scans and look for
similarities across images. It scans each person in view to see if they
correspond with the suspect’s description or photo in real time. If
there is a match, the officer is instantly alerted.
What could go wrong?
This technology could be a real life-saver in certain situations.
Despite this, AI in facial recognition is one of the debated subjects
today. Detractors think this is a major violation of individual privacy.
But the use of facial recognition is not a new concept in police work.
Police departments already use different technologies to match photos
of suspects to driver’s license photos or mug shots.
Many people are unaware of the fact that they might already be
included in a facial-recognition database of some kind. In reality, half
of American adults are already in them.
One area of concern is that the technology has proven to be somewhat flawed in the past. An MIT Media Lab study
recently demonstrated the problem. It showed that most facial
recognition technologies identify lighter faces with greater accuracy
than those with darker skin tones.
There are gender issues, too. It identifies men more accurately than
womens. This has raised red flags with human rights organizations and
privacy advocates. They are worried that these biases will lead to
misidentifying innocent people as suspects or wanted criminals.
Companies like IBM, Facebook, Google and Microsoft
are now focused on training algorithms to help this technology with
millions of publicly available photos. But at this preliminary stage,
AI-equipped body cameras remain controversial.
Axon has convened an ethics board
to help guide the development of this new technology. The board
includes civil rights activists, law enforcement officers and AI
experts, thus offering a comprehensive advisory team.
It is clear that the AI-powered face recognition technology has to
overcome these drawbacks. But that doesn’t mean its usage should be
If we are to fight crime more effectively as time goes on, then law
enforcement needs these new technologies. It is counterproductive to
stop them from using tools that can make criminals more dangerous than
About the Author
Bambi Majumdar has been writing for various industries for more
than 17 years. She has contributed articles to The Economic Times, the
leading financial daily of India, among others. She loves research,
business analysis, SEO, brand strategy and knowledge management, which
paves the way for a steep learning curve.