The Human Side of Our AI Gun Detection

by | Dec 21, 2021

Machines have been building other machines for quite some time now ⎯⎯ think about robots buildings cars on assembly lines, which began approximately 50 years ago.

Though they perform tasks autonomously, they aren’t totally self-aware; this isn’t The Terminator’s Skynet and we’re not living in the computer simulation of The Matrix. In order for AI to function and thrive, it needs to be taught and guided by humans; commonly referred to as Type II: Limited Memory Artificial Intelligence

Furthermore, for this AI solution to be effective and applicable, it needs to be trained, and this can only be done at the direction of human beings. According to the Security Industry Association’s report, 2022 for Security Megatrends, training algorithms for an AI solution “requires a massive human effort of providing nuanced feedback to the algorithm.” For more information on the different types of Artificial Intelligence and what each can do, refer to our earlier blog “What Is Artificial Intelligence, Exactly?

Only through continuous attempts at perfecting the AI and bolstering its capabilities can we offer an increasingly robust firearms detection solution ⎯ cue ZeroEyes. With hundreds of Gun Detection AI models that have been trained for over 15,000 hours, we have developed a robust, effective, and efficient solution to potential active shooter situations. Other developers attempt to attain the level of precision and accuracy that ZeroEyes has reached, but their methodology and approach simply do not yield the same results. 

What makes the ZeroEyes solution stand out from others and work better than any other product in the AI Gun Detection market, is the service that we call ‘Human-In-The-Loop.’ These are real live human beings ⎯ many of them experienced military veterans ⎯ not machines, who are trained to recognize and differentiate objects as “gun or not a gun” and respond accordingly.

Train the Brain: Our AI Training Process

 

In addition to our majority military-trained analysts who vigilantly monitor for live gun detections in our 24/7 Operations Center, we at ZeroEyes have built an entire green-screen lab to train and upgrade our AI models, which helps our model learn and remember greater varieties of weapons, clothes, colors, lighting, and varying environments. 

In the words of our Senior Vice President of Research & Development Marcus Day, because of the AI Lab, “we are able to successfully perform live demos in novel environments,” which then allows us to collect more data and record more robust data sets. These demonstrations are acted out by ZeroEyes personnel who stage as active shooters while using a diverse range of “trainer” guns.

The current ZeroEyes inventory consists of over 70 trainer guns, which include varying models of common handguns from Glock, Sig Sauer, and Smith and Wesson, to Springfield and the popular 1911 (in both 9mm and .45 caliber variations). 

Additionally, we utilize multiple 12-gauge and 20-gauge shotgun styles that include pump, semi-automatic, and break-action shotguns ⎯ and even a magazine-fed model ⎯ along with a wide variety of rifle platforms, AR pistols, PDWs (Personal Defense Weapons), submachine guns, and sniper rifles.

ZeroEyes prides itself on continuous improvement and seeking out betterment wherever and whenever possible, which we achieve by utilizing an ever-growing inventory of trainer guns paired with the technological expertise and creativity of our personnel. Through our focus, dedication, and intelligence (both human and artificial) ZeroEyes will continue to strive towards Stopping Active Shooters at First Sight, Not First Shot.

About Adam

Adam Oehrle, an Operations Center Analyst and Drone Pilot for ZeroEyes, is a veteran who served four years active duty in the United States Air Force. During his time in the military, he completed an overseas tour and served as a Command and Control Battle Management Operations specialist under the Air Force’s Central Command (AFCENT) and the United States Central Command (USCENTCOM). After returning home, he completed various schooling that led to an associate’s degree focused on criminal justice and homeland security. Furthering his education, Adam obtained his Bachelors in Criminology; focused in Politics, and Sociology.

More to Explore

Acoustic Gunshot Detection vs. Visual Gun Detection

Acoustic Gunshot Detection vs. Visual Gun Detection

As the old adage goes, don’t put all of your eggs in one basket. This is true in many cases, especially when it comes to security planning. Consider security in earlier times - it was incredibly elementary and focused primarily on physical security (usually only one...

read more
Avoiding Deep Learning AI Bias Part 2

Avoiding Deep Learning AI Bias Part 2

In Part 1 of this series, we discussed sample bias, measurement bias, and prejudice bias in Artificial Intelligence. From the beginning, ZeroEyes has set out to eliminate AI bias from our gun detection platform by focusing on object detection (in this case, weapons of...

read more
Avoiding Deep Learning AI Bias Part 1

Avoiding Deep Learning AI Bias Part 1

Sample, Measurement, and Prejudice Bias Can software be biased? We tend to think of machine and deep learning AI as consistent, logical, and unwavering, but surprisingly, that isn’t always the case. Bias is the source of many AI failures. So why, and how, does bias...

read more