That’s the goal of Wave Welcome, a software company developing real-time firearm detection technology. When given access to public or private security cameras, the AI system visually identifies weapons and moves to alert police. The company then deploys drones to follow an active shooter. It intends to reduce the number and intensity of active shootings in the U.S., according to founder and CEO Vennard Wright. 

Founded in 2020, Wave Welcome is an IT company at its core. It makes money from managing the internal IT systems of businesses, universities, and government organizations, then invests profits into its own side projects. PerVista, the gun detection tech, is one of those projects. It is available for commercial use in the Washington, D.C., and Baltimore regions and will expand across the U.S. next year. 

Located in Prince George’s County, Maryland, Wave Welcome is about 20 miles south of Washington. The county is 64% Black and 20% Hispanic or Latino, a big difference from the U.S. as a whole, where 59% of people are white, according to 2020 Census data. For decades, it was rated the wealthiest majority-Black county in America—a title it lost to neighboring Charles County only last year. Wright, a Black man, has spent the majority of his life as a resident of Prince George’s County. It is precisely his racial identity and his desire to build up his local community that led him to where he is today, he told Fortune.  

Black founders have long been underrepresented in business and in tech. While Black employees make up 12% of the U.S. workforce, they only account for 8% of jobs in tech and 3% of tech executives in the C-suite, according to a McKinsey & Company report. A measly 1% of U.S. venture capital funds went to Black founders last year, TechCrunch reported. With PerVista, Wright stands out as one of the few Black founders in the AI space and someone who is also addressing the issue of gun violence that disproportionately impacts minority communities. 

Wright began working on the technology after an incident of gun violence hit close to home. In May, three teens attempted to shoot a 14-year-old boy on a school bus in Prince George’s County. The event sparked local conversation about how to address similar situations in the future.

“I take a lot of pride in improving the area,” Wright said. “I could have founded the company right across the bridge in Arlington or somewhere else, but the fact that I’m here gives me enthusiasm.” 

The difference between a rifle and a camera tripod

PerVista’s proprietary AI works by monitoring IP-enabled security cameras and continually searching for visual matches to firearms. If the technology identifies a weapon, it automatically sends SMS texts, email notifications, and video footage to Wave Welcome and the client’s predesignated parties—security officers and teachers at a school, for example. Wave Welcome personnel who work around the clock visually confirm the presence of a firearm through the video stream before alerting police, which works to ensure emergency services aren’t being dispatched for false positives. Organizations who have paid for the PerVista service will already have two drones on-site—one inside and one outside—to monitor the situation and report to police. Armed security guards can be coded in as exceptions. 

In its testing, PerVista correctly identified 100% of weapons that appeared on camera, with a consistent confidence level of 90%, depending on the lighting or physical barriers obstructing the view, Wright said. When the system flags a false positive—a black crutch or camera tripod, for example—the company retrains the model to correctly identify those objects in the future. PerVista hasn’t flagged something that isn’t a weapon in “a couple of months,” Wright said. 

With school shootings in the U.S. trending upwards, schools are a clear market for Wave Welcome’s technology. But public schools are also notoriously underfunded. Many schools around the country are experiencing infrastructure issues, burnt-out teachers, and cockroach infestations, a CNN investigation found. Administrators are left with the choice on how to spend a limited budget, and AI gun detection systems may not make it to the top of the list. Wright recommends schools lean on government grants for public safety and education, like funds from the American Rescue Plan and Bipartisan Infrastructure Law. 

Other target markets include office buildings, sports venues, concerts, and shopping malls. “Anywhere there are a lot of people in one place, the technology would be relevant,” Wright told Fortune. 

Surveillance comes with privacy tradeoffs and critics

The project, which comes as mass shootings in the U.S. have reached a grim new record this year, is part of a new crop of AI technologies emerging to address the problem. Companies like ZeroEyes, Scylla, and Kogniz have the same basic framework as Wave Welcome, by using AI in camera systems to detect weapons and alert police. But the technology isn’t without its critics.  

The proliferation of surveillance systems could infringe on individuals’ rights to privacy, according to Scott White, director of George Washington University’s cybersecurity program. While the Wave Welcome founder said its cameras don’t start recording until a firearm is detected, that is a guardrail the company implemented. Any company with similar technology could remove that limitation with pressure from clients or in some future iteration of the technology, White told Fortune.

“You can extend the technology to the nth degree,” he said.

Other companies are already working on AI technology that can detect anomalous behavior, including suspicious shopping activity that could result in theft. Coupled with facial recognition software, police can more easily track people who break the law, but the expense to private citizens who also appear on camera is an ongoing conversation. That facial recognition tech has been criticized for discriminating against Black people is also a concern. And with such technology, a person with a lawful, concealed carry license could be flagged as a threat if the camera catches a glimpse of the weapon in a “jacket blows open” situation. Another potential privacy-defying application is identifying bulges in pockets that look like firearms, White said. 

Whether gun detection technology can actually minimize the loss of life in a significant way is another open question. A key selling point is that the tech can quickly alert police and provide a live video feed. But active shooter incidents typically end within minutes, sometimes sooner than police arrive on scene. While recent data on the timing of shootings is limited, the FBI reported in 2014 that 69% of shootings ended in less than five minutes. Response time from law enforcement varies, but 58% of shootings ended before police arrived, most often from the attacker fleeing the scene or dying by suicide, a New York Times investigation found last year.

These technologies don’t solve the underlying problem that drives people to commit mass acts of violence, White said. Individuals who engage in shootings overwhelmingly share similar characteristics, including mental health issues, social exclusion, anger management problems, and access to firearms, according to the FBI and two professors who studied the life stories of 180 perpetrators. 

“I think it’s more imperative as a society we fix those issues, and the gun issue will take care of itself,” he said. 

In Philadelphia—where data shows that arguments are the leading motive for shootings—nonprofits are teaching students conflict mediation techniques intended to reduce gun violence. Greater access to mental health screenings, resources, and guidance counselors have the potential to mitigate violence at schools, research from the American Counseling Association suggests. And regulating how guns are stored, carried, and used could reduce the number of deaths by firearms, reported the Rand Corporation, a nonprofit research group. 

While any life saved is meaningful, it will come down to how governments, companies, schools, and other organizations choose to spend their money.