Flock CEO says its microphones detect ‘just gunshots’ — its website says different

Flock co-founder and CEO Garrett Langley appeared onstage at TechCrunch Disrupt, where he defended the surveillance company's audio-collection product, called Gunshot Detection, against criticism. Asked about the gap between his description of the product and Flock's own marketing, he said flatly, "It's just gunshots." The company's website describes considerably more capability, saying the audio systems can pinpoint gunfire instantly while also detecting street takeovers early, identifying severe crashes in real time, and sending alerts for car crashes, fireworks and sideshows.
The feature was announced last year and drew immediate pushback. At launch, Flock said the technology would listen for sounds of "screaming." Privacy groups condemned the company for listening for human voices, after which Flock amended its website wording to "sounds of human distress." The company later said it pulled the ability to detect "distress" altogether from the customers who had tried it. A Flock spokesperson, Paris Lewbel, told TechCrunch that the gunshot detection technology is primarily used to detect gunfire but is also capable of detecting fireworks, sideshows and potential car crashes, and that Langley's remarks were intended to emphasize that the Audio Detection technology is not designed to monitor conversations or analyze speech and does not currently detect human screaming or distress. An FAQ on the site states that in addition to gunshots, Flock Gunshot and Audio Detection can detect fireworks, sideshows and other disruptive community events, and that its detection library is continually expanding to cover emerging threats relevant to customers.
Langley framed the origin of the product around emergency response, telling TechCrunch editor-in-chief Connie Loizos that most people do not call 911 when they hear a gunshot, which is why the company built it. He described how some customers who had the devices in parking lots asked to be alerted if someone was being abducted or harmed. He said the idea seemed logical, so Flock prototyped it and rolled it out to a couple of customers, and that some people then raised many concerns. He conceded the company made mistakes but said it was listening to its customers. In his account, only a couple of customers were prototyping the broader capability before the company narrowed the product, telling them it would "just... do gunshots, plain and simple" — notifying when a gun goes off and listening to nothing else.
On scale, Langley said the number of Flock microphones is comparable to ShotSpotter, which per Wired reporting had about 25,000 microphones as of 2024. That is the only sizing figure he offered. Flock does not disclose an exact microphone count in the remarks, so the comparison is best read as an order-of-magnitude claim rather than a precise tally.
The appearance came amid mounting scrutiny of Flock from privacy advocates including the ACLU and from members of the public worried about misuse of its technology, with some going as far as destroying the surveillance cameras. Texas and Florida have moved to end their relationships with Flock, and Amazon's Ring has ended its partnership with the company. Protesters gathered outside the San Francisco venue as Langley took the stage, and he used the interview to push back on what he described as misinformation, arguing that many critics misunderstand how the products work.
The episode is a familiar pattern in surveillance technology: a capability is marketed broadly to public agencies and property owners, then narrowed after advocacy pressure, leaving a gap between what the hardware is described as doing and what executives say it does. This suggests the dispute is less about whether the microphones exist than about what they are trained to classify. Flock's FAQ language about a continually expanding detection library implies the set of recognized sounds is not fixed, which in practice makes any "just gunshots" characterization a statement about current configuration rather than a hardware limit. A likely trade-off is that expanding the classifier to crashes, fireworks and sideshows improves coverage for customers but widens the surface for the privacy objections that have already forced two rounds of public wording changes.
Ambient-audio detection systems typically work by classifying acoustic signatures on or near the device rather than recording and transcribing speech, which is the distinction Flock leans on when it says the product is not designed to monitor conversations or analyze speech. Whether that distinction satisfies critics is a separate question the company has already had to answer twice.
Why it matters: Any developer or product team building or buying into municipal sensor networks should treat marketing pages and executive statements as separate signals, because the two can diverge in ways that create legal and reputational exposure. Procurement teams evaluating Flock now have contradictory public descriptions of the same product, and the state-level pullbacks in Texas and Florida plus Ring's exit show that contracts can unwind quickly when that gap becomes public. For teams designing audio or sensor pipelines more generally, the case is a reminder that a continually expanding detection library needs clear, versioned statements of what a system will and will not classify, since inference about future capability will be drawn from today's marketing copy.
Based on reporting from the original publisher. Visit the source for full context and later updates.
Publisher excerpt
Langley’s comments come amid mounting scrutiny of Flock from privacy advocates, including the ACLU, and members of the public who are concerned about the misuse of its technology.