CNN  — 

The number of shootings involving automatic weapons in the US has skyrocketed in recent years, suggesting a troubling resurgence in the use of machine guns by criminals, according to an analysis provided exclusively to CNN.

Incidents of machine gun fire have exploded by about 1,400% from 2019 through last year, according to statistics compiled by a gunfire detection company that has acoustic sensors placed in about 130 US cities. Last year alone, ShotSpotter, Inc. detected roughly 5,600 incidents of automatic weapons fire, the analysis showed.

The previously unreported figures add to growing evidence that the widespread availability of inexpensive so-called conversion devices – known as “auto switches” or “auto sears” – capable of transforming semi-automatic weapons into machine guns in a matter of moments are wreaking havoc on American streets.

There has been a corresponding spike in seizures of conversion devices by the Bureau of Alcohol, Tobacco, Firearms and Explosives in recent years, from fewer than 100 in 2017 to more than 1,500 last year.

CNN earlier reported on an ATF effort in 2019 to recover what agents suspected were thousands of such devices illegally imported from China. In the years since, 3D printers have only added to the problem, law enforcement officials said.

“Not since Prohibition have we seen this many machine guns being used to commit crimes,” said Tom Chittum, who spent more than two decades with the ATF and retired as its No. 2 official before signing on as an executive with ShotSpotter earlier this year.

Gun laws virtually eliminated automatic weapons from city streets for decades, Chittum said. “But now machine guns are back, and they’re everywhere.”

Automatic weapons have been used in several high-profile assaults in recent years, including a mass shooting in Sacramento, California, in which six people were killed and a dozen injured, a school shooting in Washington, DC, in which a sniper with automatic rifles unleashed a barrage of more than 200 shots, wounding four, and the slaying of Houston police officer William “Bill” Jeffrey during the service of an arrest warrant last year. The 30-year-veteran officer died in a torrent of gunfire from a convicted felon armed with an illegally converted weapon.

The scene of a mass shooting is blocked off with police tape on April 3, 2022 in Sacramento, California.

A CNN review of court filings in cities across the US found dozens of cases in recent years involving so-called conversion devices or semi-automatic handguns already converted to fully automatic.

In Chicago, a man prosecutors called “a prolific machine gun dealer” allegedly continued to sell the devices while out on bond and awaiting trial. An alleged associate of the man was recorded telling an undercover ATF agent posing as a buyer that he’d get a better price if he bought in bulk, and that he should act quickly because demand was high.

“People are gonna get them switches,” he told the agent, according to a court filing. “It’s gonna go to the people who want to go shoot some people, gangbangers and sh*t.”

The devices appear to be an emerging commodity on the black market. CNN reviewed cases in which they were allegedly hawked on social media, sold under the table by a licensed gun dealer in Miami, and turned up in the possession of alleged drug dealers distributing methamphetamine, fentanyl and oxycodone. In Los Angeles, a man under investigation for supplying local gang members with guns allegedly sold an ATF informant a Glock conversion device along with a Glock 9mm pistol and a high-capacity magazine. In Washington, DC, investigators looking into a young man who allegedly tossed a converted handgun into a trash can as police approached later found several videos on YouTube in which he rapped about “switches.”

The increasing availability of auto switches has been driven in part by the ease with which they can be made using cheap, 3D-printed parts and instructions available online, according to Earl Griffith, the chief of ATF’s Firearms and Ammunition Technology Division.

“It’s very easy,” said Griffith, who explained how he learned to use a 3D printer to make the devices on YouTube. “In a matter of 15 minutes I was able to do it myself the first time.” Below, Griffith explains.

A CNN review of YouTube based on key-word search terms revealed multiple such videos that had collectively racked up more than 1 million views. One group of auto-switch instructional videos that remained online until August were linked to a man charged in December by federal prosecutors in Texas for allegedly making, possessing and transferring 3D-printed switches. He has pleaded not guilty.

YouTube removed the videos after CNN asked about them. A company spokesperson said YouTube does not allow “content instructing viewers how to manufacture accessories that convert a firearm to automatic fire, or to sell those accessories on our platform.”

Griffith said that despite the growing ubiquity of the devices, many members of law enforcement do not know how to recognize them on firearms they seize from criminals.

“When we tell them about it, they go back into their evidence vault and they look and check and they find this stuff,” he said.

Quantifying incidents involving automatic weapons fire is a challenge. The shell casings fired by automatic weapons appear no different than those discharged from a semi-automatic gun.

That is where ShotSpotter comes in. The company has contracts with about 130 cities in which it installs acoustic sensors in designated areas to listen for gunfire. A patented computer algorithm attempts to distinguish between innocuous sounds such a jackhammer or car backfiring and gunfire. Human analysts at the company’s headquarters in the San Francisco Bay Area or a satellite office in Washington, DC, listen to what the algorithm flags as gunfire and, if they concur, alert police. The company says its goal is to make such notifications within a minute.

In recent years, suspected incidents of automatic weapons fire have risen sharply from about 400 in 2019 to 1,800 in 2020 to 5,600 last year. Even after adjusting for an increase in the company’s coverage area in the US, ShotSpotter said an internal analysis showed that suspected automatic gunfire incidents jumped 14 times in about three years. The upward trend has continued in the first half of this year with roughly 3,800 incidents detected. The company says its designation of an incident as “full auto” is for a police department’s “situational awareness” only and is not guaranteed in the same manner as its primary mission of accurately identifying and locating outdoor gunfire.

During a demonstration of the system in June, analyst Kaylan Parker replayed some of the incidents she and others had tagged as “full auto,” filling her Washington, DC, listening post with audio from what sounded like some far-off war zone. On a recent day, she said, the company had detected what it determined to be more than 25 incidents of automatic weapons fire, involving some 300 rounds, including a shooting in nearby Baltimore. Baltimore police later issued a press release about the incident, citing the ShotSpotter alert and stating that two people were wounded, including a 14-year-old boy.

Founded in 1996, ShotSpotter bills itself as an important tool for police, providing real-time information about the location and nature of shootings, which the company says often go unreported. The early intel, company officials say, provides a tactical advantage to police and has resulted in both the arrest of shooters and faster medical care for gunshot victims.

But ShotSpotter, a publicly traded company with reported revenue last year of nearly $60 million, has been mired in controversy in recent years. The criticism is centered on the placement of its sensors in predominately minority communities and the use of its information as evidence in court cases as opposed to its primary mission of merely alerting police to the occurrence and location of gunfire. Critics see the placement of the sensors as racially biased, resulting in the increased use of stop-and-frisk tactics by police. Defense attorneys have assailed ShotSpotter’s results as both unreliable and impossible to scrutinize because the company has declined to disclose the precise science behind how its system works. Other critics have questioned ShotSpotter’s true value as a crime-fighting tool, regardless of how well it detects and locates gunfire, because they say there’s no compelling evidence that it reduces gun violence.

ShotSpotter, which touts a 97% accuracy rate – a figure backed by an audit paid for by the company – has pushed back against the criticism with a prominent link on its website. It cites studies it says are “proof of its positive impact” and says the location of its sensors is decided in consultation with police and city officials in the communities it serves and is “based on historical gunfire and homicide data.”

A case winding its way through federal court in Washington, DC, highlights both the utility of ShotSpotter, and the challenges prosecutors sometimes face when trying to use its information as evidence in court.

Early on the morning of January 20, 2020, ShotSpotter notified police in Washington, DC, of gunfire at a house in the city’s southeast quadrant. Police later discovered that footage from a surveillance camera mounted nearby showed a man firing a weapon into the air at 4:45 a.m., the precise time of the ShotSpotter alert, according to court records.

After obtaining a search warrant, police found a twice-convicted PCP dealer alone in the house. They also seized a Glock .40 caliber semi-automatic handgun equipped with a conversion device, and an extended magazine, from a closet, according to prosecutors. The occupant of the house was arrested and charged with possession of a machine gun.

It might appear an open-and-shut case for the efficacy of ShotSpotter. But the use of the company’s information as evidence in court has been another matter.

Prosecutors and defense attorneys in the case have been battling for months over who is qualified to provide expert testimony regarding ShotSpotter’s findings and address questions big and small about the company, from the science behind how its system works to an explanation of how the estimate of the number of shots fired changed over time in the case at hand.

As of publication, a judge had yet to rule on how the ShotSpotter information would be handled.

In Texas, Lacie Jeffrey, the daughter of the Houston police officer killed last year, said she could not comprehend the proliferation of fully automatic weapons like the one that riddled her father’s body with multiple gunshot wounds in an instant. An autopsy report obtained by CNN shows that the veteran officer was struck more than a dozen times during the brief encounter.

Jeffrey said she has reached out to lawmakers in Texas in hopes of enacting a tougher state law regarding so-called conversion devices like the one on the weapon used to kill her dad.

“We do not live in a war zone,” Jeffrey told CNN. “There is no need for us to have these automatic weapons on the streets of Houston — anywhere in the United States.”