More bad news for ShotSpotter, which recently re-branded to “SoundThinking” to distance itself from exactly this sort of negative press. Four legislators (three senators, one congressperson) are asking the DHS Inspector General to take a closer look at the tech the DHS is funding via one of its grant programs.

The problem with Spotshotter is it seems unlikely to put a dent in the public’s arsenal. Multiple cities have chosen to dump the tech rather than continue to pay for false positives, altered shot reports, and nonexistent public safety increases.

The problem with the DHS is that it has already started spending money on a portable “Gunshot Detection System.” It’s capitalized for a reason. It’s a bespoke version of a product already offered by a company called [re-reads DHS press release] Shooter Detection Systems — a redesign of its [deep breath] Guardian Indoor Active Shooter Detection System.

According to the DHS’s PR team, the “enhanced” version of this off-the-shelf shot spotter will detect both sounds and light flashes, apparently aiming to reduce the number of false positives generated by acoustic-only detection systems… like the one offered by [coughs at first half of rebrand] SoundThinking, formerly ShotSpotter.

Whether adding “eyes” to “ears” to spot shots has accomplished a reduction in false positives is still an open question. Whether or not the DHS should continue to pay for shot spotting tech — namely the one offered by the former ShotSpotter — is exactly the question these four lawmakers would like the DHS Inspector General to answer.

The question — as posed in this letter [PDF] from Sen. Ed Markey, Sen. Ron Wyden, Sen. Elizabeth Warren, and Congressperson Ayanna Pressley — is a bit leading perhaps. But the question is valid and the lawmakers’ letter contains plenty of evidence that lends validity to the question: should the DHS really be spending federal dollars on grants to local law enforcement agencies seeking to acquire ShotSpotter tech?

Several recent reports have cast substantial doubt on the accuracy and effectiveness of the “ShotSpotter” gunshot detection system and have raised serious questions about its contribution to unjustified surveillance and over-policing of Black, Brown, and Latino communities. Through the Urban Area Security Initiative (UASI) grant program, the Department of Homeland Security (DHS) provides funding to localities to deploy the ShotSpotter system. We request that the DHS Office of Inspector General (OIG) investigate DHS’s spending of taxpayer dollars on ShotSpotter, including potential violations of Title VI of the Civil Rights Act of 1964, which prohibits recipients of federal financial assistance from discriminating based on race, color, and national origin.

And that’s only part of the problem. It’s not even necessarily a ShotSpotter problem per se, but a long-standing problem with law enforcement agencies, who almost always deploy new surveillance “solutions” in low-income neighborhoods, especially those heavily populated by minorities.

The other problem is more technical: the tech just doesn’t work as advertised. Multiple investigations have shown the tech is either (1) unable to reliably detect gunshots, or (2) doesn’t lead to better enforcement of gun-related crime. The cities now dumping the tech say it’s both unreliable and useless. Of course, SoundThinking/ShotSpotter insists otherwise in responses to the latest negative reporting and in its marketing materials, which are still somehow capable of convincing government entities to buy its tech.

That’s where the DHS comes in. It offers grant money to law enforcement agencies — funding that can be used to purchase acoustic gunshot detection tech. The biggest brand in the business is SoundThinking, so naturally that’s where most of this funding goes.

In Massachusetts alone, “UASI [Urban Area Security Initiative] has funded almost a decade of contracts for gunshot detection technology with ShotSpotter in Cambridge, Chelsea, Somerville, and Boston.” Since 2012, according to city records, Boston has spent more than $4 million on ShotSpotter. Elsewhere, municipalities across the country have used UASI funds for the ShotSpotter system. One study found that “[t]hrough an analysis of UASI funding in Los Angeles, Boston, New York City, and Chicago . . . cities spend millions of UASI dollars on contracts with surveillance corporations” such as ShotSpotter.

And what are we getting in return for this combination of federal and local spending? Not much.

The ShotSpotter system’s ineffectiveness has consequences for law enforcement, community response, and the prevention of gun violence. A 2021 study from the Journal of Public Health found “that implementing ShotSpotter technology has no significant impact on firearm-related homicides or arrest outcomes” and that “[p]olicy solutions may represent a more cost-effective measure to reduce urban firearm violence.” Another study from the MacArthur Justice Center at Northwestern University concluded “that more than 90% of ShotSpotter alerts lead police to find no evidence to corroborate gunfire when police arrive at the location ShotSpotter sent them: no shooting, no shell casings, no victims, no witnesses, no guns recovered.

Not great. More bad stuff from studies and reports: 70% of people in neighborhoods with ShotSpotter systems are either black or Latino. 75% of those neighborhoods had annual incomes well below the national median.

As I stated above, this is a cop problem: the long-held biases that subject the same people to any new surveillance option. The rest of it is a ShotSpotter problem: it doesn’t spot shots and it doesn’t stop crime. And yet, millions are being spent on it every year, with some of the funding flowing directly from the DHS.

The main point of this letter, however, is to nudge DHS oversight to take a close look at the end result of this funding in terms of purchasing ShotSpotter tech. The federal government is forbidden from spending money on anything that violates federal laws. And this funding might be doing that. Title VI of the Civil Rights Act of 1964 forbids recipients of federal funding from discriminating on the basis of race, color, or national origin. Does planting most of your shot-spotting mics in predominately non-white neighborhoods violate the Civil Rights Act?

Well, that’s what these lawmakers hope to find out.

For all the preceding reasons, we respectfully request that you open an investigation in DHS’s funding of the ShotSpotter system to determine whether it is an appropriate use of taxpayer dollars, including the critical question of whether such funding may lead to Title VI violations.

It will likely be awhile before we hear back on this. But given what’s already been discovered via studies, public records requests, and investigative journalism, it certainly looks as though cops with this tech are violating the law. And one would expect another investigation into ShotSpotter use is going to turn up more of the same biased policing. If that’s the case, it won’t stop cops from being racist. But it will mean they’ll have to spend local funds to keep minorities under their tech-enhanced thumbs.

Sole source contracting is the sort of thing government agencies should seek to avoid. In some cases, it’s impossible, but most spending should be open to bidding to help ensure the government isn’t spending more than it has to — or worse, hooking up contractor buddies Mob-style.

There’s a whiff of impropriety in all of this, but it may be imaginary. However, it’s still worth examining more closely, as the Associated Press has done here. School shootings just aren’t going to go away here in the United States, so it’s understandable that legislators and educators are exploring their options. But what’s detailed in this article suggests several things, most of which aren’t exactly great.

Kansas could soon offer up to $5 million in grants for schools to outfit surveillance cameras with artificial intelligence systems that can spot people carrying guns. But the governor needs to approve the expenditures and the schools must meet some very specific criteria.

The AI software must be patented, “designated as qualified anti-terrorism technology,” in compliance with certain security industry standards, already in use in at least 30 states and capable of detecting “three broad firearm classifications with a minimum of 300 subclassifications” and “at least 2,000 permutations,” among other things.

Only one company currently meets all those criteria: the same organization that touted them to Kansas lawmakers crafting the state budget. That company, ZeroEyes, is a rapidly growing firm founded by military veterans after the fatal shooting at Marjory Stoneman Douglas High School in Florida.

The first thing worth closer observation is this part of the AP’s reporting: “Only one company currently meets all those criteria.” That company would be ZeroEyes, which has benefited from similar legislation and similar grants across the country.

Again, it’s understandable that if the goal is protecting children, legislators and schools are going to want the best product. At this point, that would appear to be ZeroEyes. It does have competitors but almost none of them are able to meet the very specific criteria that keep showing up in new gun detection tech grant legislation.

ZeroEyes also appears to be the only firm qualified for state firearms detection programs under laws enacted last year in Michigan and Utah, bills passed earlier this year in Florida and Iowa and legislation proposed in Colorado, Louisiana and Wisconsin.

On Friday, Missouri became the latest state to pass legislation geared toward ZeroEyes, offering $2.5 million in matching grants for schools to buy firearms detection software designated as “qualified anti-terrorism technology.”

This sort of thing would appear to be outside the boundary of normal coincidence. And yet, the co-founder of ZeroEyes told AP “We’re not paying legislators to write us into their bills.” That’s probably true. The days of garishly showing up at a lawmaker’s office with a briefcase full of money are (mostly) behind us. But the ability to influence not just the direction of legislation, but the creation of legislation is something industries and their lobbyists have been capable of accomplishing without having to actually “pay” anyone to do anything.

Now, it could be that ZeroEyes just got out in front of its competitors to meet a bunch of requirements that its competitors couldn’t. Or it could be that it chose to approach regulators, rather than legislators, to help ensure the specifications and requirements more closely matched its product than any of its rivals.

Or, it’s just a coincidence ZeroEyes is seemingly the sole beneficiary of recently crafted legislation that provides funding to schools to buy tech from a single source.

Given all of this, it would seem most likely ZeroEyes is persuading legislators these laws need to be written and these funds need to be available to protect children from school shootings. That only its product meets the requirements is a happy coincidence, rather than the result of loutishly dangling campaign contributions over malleable legislators’ heads.

America is a business more than it’s a country. These new laws are now forcing taxpayers to fund tech that hasn’t exactly shown it can handle the responsibilities expected of it. Given the numerous options available to legislators, law enforcement, and a small army of government agencies (including social services and mental health professionals), is it really the best idea to start throwing money at the shiniest option, especially when only a single provider meets the criteria… which means it doesn’t need to be all that shiny to begin with?

Here’s what the chairperson of the National Council of School Safety Directors (Jason Stoddard) had to say about ZeroEyes and its legislation-enabled takeover of the AI gun detection market:

The super-specific Kansas bill — particularly the requirement that a company have its product in at least 30 states — is “probably the most egregious thing that I have ever read” in legislation…

[…]

When states allot millions of dollars for certain products, it often leaves less money for other important school safety efforts, such as electronic door locks, shatter-resistant windows, communication systems and security staff, he said.

“The artificial-intelligence-driven weapons detection is absolutely wonderful,” Stoddard said. “But it’s probably not the priority that 95% of the schools in the United States need right now.”

That’s probably the most salient point, whether or not ZeroEyes’ hands are completely clean. Most AI in use today is still a bit wonky. Facial recognition tech has been around for years, but most products still suffer from the same issues, even though they’ve had plenty of time to address things like built-in bias or, you know, just assuming any images they find laying around the ‘net are free for the taking (and training).

And as for the co-founder’s claims it does not buy legislators or legislation, it hardly seems to matter. Purchased or not, legislators are willing to make ZeroEyes a sole source for in-school gun detection tech. As the AP reports, the company put on a presentation for state legislators earlier this year. Whatever else went on between ZeroEyes and Kansas legislators, this was (almost!) the end result:

Kansas state Rep. Adam Thomas, a Republican, initially proposed to specifically name ZeroEyes in the funding legislation. The final version removed the company’s name but kept the criteria that essentially limits it to ZeroEyes.

Yeah, that’s not a great look, even if ZeroEyes stayed out of the law-crafting process. Sadly, another Republican on the K-12 budget committee was even more voracious in the defense of ZeroEyes as a sole source provider, telling her fellow legislators the state “couldn’t afford the delays of a standard bidding process.” Whew.

All of this may just be coincidence. But given the reach of ZeroEyes and the number of states that have passed similar legislation that solely benefits the only company that (equally coincidentally!) can match specific criteria enumerated in these bills, there’s reason to doubt every bit of this is on the up-and-up.