Flock's 'just gunshots' claim runs into its broader audio-detection pitch
After its CEO's Disrupt remarks, Flock confirms detection of fireworks, sideshows and potential crashes while denying conversation monitoring and speech analysis.
A narrower stage claim meets broader product copy
Flock Safety’s audio-surveillance pitch is facing a basic question: what do its microphones actually detect? At TechCrunch Disrupt on October 14, CEO Garrett Langley repeatedly described the product as detecting “just gunshots,” according to TechCrunch. The company’s English product page advertises gunfire, fireworks, severe crashes and sideshows, including street takeovers.
Asked by TechCrunch about that gap, spokesperson Paris Lewbel confirmed that the technology is primarily used for gunfire but can also detect fireworks, sideshows and potential car crashes. Lewbel said Langley’s remarks were meant to emphasize that the system is not designed to monitor conversations or analyze speech, and does not currently detect screaming or human distress.
Those are different boundaries: recognizing several categories of sound is broader than detecting gunfire alone, but it does not by itself establish that the product understands conversations.
Processing sound and retaining clips are separate questions
Flock’s own explanatory page says sensors continually analyze short rolling windows on the device and retain encrypted clips when a qualifying event occurs, with a separate allowance for limited scheduled diagnostic samples. The company says there is no live listening feed and that it does not transcribe speech, identify voices or let customers search spoken words.
The same explanation acknowledges that an event clip can incidentally contain a few nearby spoken words. That qualification matters when assessing a microphone deployed in public space: a denial of speech analysis is not a promise that recorded event clips can never contain speech.
TechCrunch also reports that Flock removed an earlier, limited human-distress detection pilot after concerns were raised. The present news is the October 14 exchange and clarification about existing capabilities. The company pages describe its design and safeguards; they do not provide an independent audit of every deployed sensor’s accuracy or privacy behavior.