Two weeks before the RTIC vote, the Policy, Finance and Infrastructure Committee heard the case for it: from Axon and Flock sales representatives, presenting alongside APD, with no public comment permitted. This was the “full conversation” the April 14 delay bought. The full video runs just under two hours; these are the load-bearing claims, timestamped, checkable, each set against the documented record. (CAMS’s fact-check one-pager responded to this same session; captions don’t label speakers, so attributions below follow the presentation context.)
”Flock’s cloud environment has never been hacked. It has never been breached.” (1:42:09)
Stated flatly, twice, by the vendor side after questions about the widely reported camera hacks, which were waved off as device-level misconfiguration.
The record: the “cloud” qualifier is what makes the sentence technically true. Around the “cloud,” the documented security failures stack: researchers rooted a Flock camera in under 30 seconds; ~70 Flock cameras were found streaming unauthenticated on the open internet; Senator Wyden and Rep. Krishnamoorthi asked the FTC to investigate, citing no mandatory two-factor authentication and at least 35 cases of stolen Flock customer passwords; stolen credentials get into the “never-breached” cloud without any breach. (The National Wave) Narrow denials delivered as blanket assurance are precisely the pattern that got Flock terminated in Sedona (“misled and lied to”) and rescinded in Oshkosh within 48 hours (How Cities Won).
The immigration filter: “That search… will never hit your cameras.” (1:46:20)
“Chief, you talked about immigration. That is not an acceptable use in Asheville. So y’all actually, through our team, put the immigration filter on your Flock cameras… if [outside] agencies who are sharing your cameras conducted a search that was not allowed by your policy on immigration, it will never hit your cameras. That search is filtered out.”
Also, at (1:38:23): “we do not use or track or report immigration status.”
The record: the filter screens the reason category the searching officer selects. The Texas deputy who searched 83,000 cameras over an abortion had the search relabeled as a “missing person” case; the ACLU documented officers coaching each other to stop typing honest reasons once audits drew attention; 4,000+ immigration lookups nationally were run by local officers as favors, many under vague labels. The filter only screens the reason the officer chooses to give. It catches honest officers; a dishonest label sails through. (The Abuse Record · Their Claims vs The Record)
Facial recognition: “Fusus will not pull in the facial recognition part of that camera” (1:32:11); and APD: “We’ve never had a contract with Clearview… we do not use facial recognition” (1:50:13)
A council member relayed community concern that APD might already hold facial-recognition contracts; APD denied any Clearview contract or facial-recognition use.
The record: the denial is now on tape and testable: a records request for any APD contracts or trial accounts with facial-recognition vendors (Clearview, PimEyes, Rekor, or Axon’s own offerings) either confirms it or doesn’t; that goes on the Documents request list. The structural point stands regardless: no ban exists, and the platform vendor (Axon owns Fusus) controls the feature roadmap for the next 7.5 years. (Promise Tracker row 1)
Retention: “Flock is 30. Axon is 90 days.” (1:39:53)
The record: the public messaging has emphasized the 30-day figure. On the vendors’ own account at this session, that’s Flock only: the Axon side of the system retains for 90 days, the statutory maximum, with categories kept for years or “forever” (felony homicide) and annual re-preservation available beyond that (1:24:00). Meanwhile Flock’s own recommended default is now 7 days. When you hear “we only keep data 30 days,” ask: which system, which category, and who re-certifies the exceptions?
”We will not generally surveil or monitor for any reason. It’s in our policy.” (1:51:47)
APD’s answer to a council member’s genuinely good question. After relaying that the Housing Authority’s director privately called the surveillance “a critical tool,” the member asked: “Are we watching people sleeping in doorways? What are the limits? What will council be able to do to massage this… around what will we not watch?”
The record: that question (what can council actually control) never got a structural answer, because the honest answer is: nothing binding. The limits live in APD policy (Policy 505, revised days before the vote), revisable without a council vote. The exchange also confirms: Housing Authority properties are already integrated: ~500 of the county network’s cameras sit in housing communities and businesses, and the residents being watched were not the ones consulted. The councilmember’s instinct was right; the mechanism to act on it (an ordinance) was never offered.
”Even the vendors themselves do not have access to our data” (1:09:13)
Chief Stepp’s flat claim, early in the presentation: “even the vendors themselves do not have access to our data. We own all of that data.”
The record: contradicted twice before the session ended. Forty minutes later, Flock’s own representative described the contract’s actual terms (1:48:40): “There are two provisions… in which we are allowed to use your data. Number one, to provide the services… Number two is there is a small subset of training data that is taken… for training purposes of training the machine learning.” And Kim Roney read the operative clause into the record verbatim (1:47:33): the city “grants to Flock a limited, non-exclusive, royalty-free, irrevocable, worldwide license to use the customer data.” Irrevocable. Worldwide. The city attorney promised a legal opinion on that provision (1:48:01); whether it was ever delivered is a standing records question. The Axon side has its own version: ACEIP, the data program the contract enrolls the city in by default.
”It’s literally just brick and mortar”: same day, refuted by the vendor’s own marketing (1:06:55)
Stepp minimized: “A real-time intelligence center could be this room. It’s literally just brick and mortar. I have a computer and a person.” And on AI (1:32:11): Fusus “will not pull in AI technology.”
The record. That evening, at the council session where the public could actually speak, resident Ben Spencer read Axon’s published Fusus marketing into the record (evening video, 1:49:50): “Fusus applies AI-driven object detection and classification across connected cameras… Fusus behavioral AI enables agencies to detect motion events and activity patterns… Fusus supports integration with select third-party AI solutions.” The staff description and the vendor’s sales copy cannot both be true, and only one of them was written to close a sale to police departments. Spencer’s summary of the afternoon stands on its own: “45 minutes of interim police chief and two paid representatives for private surveillance companies flown out here does not seem like a really balanced way to gauge the facts.”
The Durham slide collapses live (1:26:15)
Staff’s peer-cities slide listed Durham’s center as “in process” via “the same funding channel.” Roney, who had done the homework staff hadn’t: “I’ve talked to three council members from Durham who said staff pulled it because it doesn’t have the council support… That is not what I’m hearing from multiple council members.” Under questioning, the vendor side conceded the Durham contract “was pulled. They declined having that conversation, and that funding was not approved”, and City Attorney Branham, live-texting Durham’s city attorney, confirmed Durham had “currently tabled” its equivalent (1:28:52). The evidence offered for the RTIC’s normalcy fell apart in the room, on camera. It’s also the second time Durham has cut the other way: Hess’s own example city rejected ShotSpotter and scaled unarmed response.
”I’m not here to sway a vote” (1:54:10)
Stepp closed the hour: “I’m not here to sway a vote, but I am here to make sure that you feel educated enough and informed enough to make a vote.” The hour in question: the chief plus two vendor representatives flown in, no opposing expert, no public comment permitted. Three more process facts from the same closing minutes: Mayor Manheimer asked “Is there a deadline with the grant?”; staff: “There’s not”; “Okay, so we have time to breathe” (1:54:58). Asked whether the item would go to the Public Safety Committee next: “No. No, I don’t think so” (1:54:37); the RTIC skipped the one committee built for it, whose chair was its loudest advocate (Oversight Theater). And Stepp herself noted the city was buying a second RTIC while “the Sheriff and I both walked away and agreed one is better” (1:04:10). No deadline, no committee review, a duplicate facility its own chief questioned, and a vote in fourteen days.
Why this page exists
This session was the process. No public comment, sales reps as the expert witnesses, claims delivered with confidence that the documented record (audits, terminations, federal investigations) contradicts or narrows to nothing. Two weeks later the motion was seconded before the public spoke. Every claim above is now on the record with a timestamp; several are testable by records request; and per How Cities Won, a vendor caught in a false statement to a council is the single fastest flipper of council votes in the country. Keep checking.
Related: Their Claims vs The Record · 2026-05-12 City Council · Promise Tracker · Documents