By: Alex Wasson, Director of Security

It may be difficult to remember when COVID wasn’t the most pressing issue on everybody’s mind. But in recent years, increases in gun violence have presented a significant challenge to the security industry. And despite nearly a year of limited social interaction, these increases have held steady.

With these concerns in mind, security manufacturers and developers are working on new tools to drastically reduce response times in the event of an active shooter.

Ideally, these solutions leverage video surveillance, powerful analytics and an AI-based engine to identify weapons before an incident occurs. Once a weapon is detected, the system automatically escalates the information along the user’s channels. Alerts can notify the building or company’s security group, the appropriate law enforcement authorities, or, as has often been proposed, a mass-notification system.

A Revolution in the Making, or Another False Start?
There’s no doubt that if this vision is realized and properly implemented, the technology could be revolutionary and save lives. However, when it comes to new security and surveillance technologies, there is a need for caution.

If this technology were to produce a false alarm, the consequences would be disastrous. Imagine if the system were to incorrectly identify an object being pulled from a backpack as a gun and trigger an alert, causing security to apprehend the individual in full view of the public.

This could cause a panic without justification. Once the truth of the situation comes to light, it could damage the reputation of the company or facility involved, as well as stoke public distrust of video surveillance and artificial intelligence.

As I discussed in a previous article about facial recognition analytics, the security industry needs to build trust with the public and demonstrate that it can be responsible with its use of analytics. Any potential application of these technologies must be carefully vetted and refined, so that mistakes are kept to an absolute minimum or eliminated.

Refinement and Implementation
In the case of weapon detection, the necessary level of refinement likely means improving the specificity of the weapon profiles that the software is capable of identifying. Currently, the software can reliably identify if an object is a knife, and has a promising success rate identifying if an object is a handgun, but promising doesn’t quite cut it.

However, if the analytic improves to a greater level of detail, it could allow for even the specific make and model of a firearm to be identified reliably. This would greatly decrease the chances of misidentifying a different object as a gun.

Then comes the question of implementation. Does a mass notification component really make sense? Given the lack of public familiarity with this technology and the high likelihood of causing a panic rather than an ideal response from most people, I would argue this should be implemented last, if at all. Beyond that, it should be done with a layer of human intervention and verification to ensure the alert is warranted.

Instead, it makes much more sense to send alerts to the building’s security group or law enforcement, giving them the ability to intervene much more quickly and potentially even prevent incidents from becoming serious in the first place.

A Constantly Advancing Industry
While it is important to emphasize caution around developments like this in the short term, that shouldn’t diminish the technology’s exciting potential. If we can clear a few obstacles, weapon detection shows clear value as a life-saving solution for the relatively near future.

Consider the state of surveillance analytics as recently as the middle of last decade. Even a simple line-crossing analytic was state of the art, and required working out a few kinks. The rate of advancement in the field of AI and analytics is truly remarkable, and it’s possible that weapon detection may be ready sooner rather than later.

However, security providers and integrators who value the long-term relationships they have built with customers will prioritize responsibility and practicality. This means setting reasonable expectations rather than pushing technologies that aren’t quite ready for everyday application yet. When the day comes, the benefits and added peace of mind will undoubtedly be worth the wait.