Imagine standing on a busy street corner, waving at a camera mounted high above, only to realize that anyone with an internet connection could be watching you in real time—no password, no login required. This is the chilling reality we uncovered when Flock, a company specializing in AI-powered surveillance cameras, left dozens of its Condor cameras exposed to the open internet. But here’s where it gets even more unsettling: these weren’t just any cameras. Unlike Flock’s more common license plate readers, the Condor cameras are designed to pan, tilt, and zoom in on people—tracking their movements, capturing their faces, and even following them as they walk through public spaces. And this is the part most people miss: anyone could access not just the live feeds, but also 30 days of archived footage, control settings, and even run diagnostics on these devices.
I found myself at the intersection of Harris Road and Young Street in Bakersfield, California, staring up at one of these cameras. On my phone, I watched myself in real time as the camera livestreamed my every move. I stepped into the crosswalk, waved at the camera, and there I was—clear as day on the feed. Hundreds of miles away, my colleagues were watching too, all through an exposed livestream that required no authentication. It wasn’t just me; we observed these cameras zooming in on a woman walking her dog on a bike path in Atlanta, tracking a man through a Macy’s parking lot, and even surveilling children playing on a swingset. In one striking instance, we watched a man rollerblading down a bike path in Brookhaven, Georgia. The camera zoomed in on him, tracked his movements, and minutes later, he appeared on another exposed livestream further down the path. The resolution was so sharp that when he stopped to watch a video on his phone, we could see exactly what he was doing.
But here’s where it gets controversial: Is this the future of public safety, or a dangerous overreach of surveillance technology? Flock’s Condor cameras are marketed as tools for security, but their exposed feeds raise serious privacy concerns. What happens when anyone—from curious onlookers to malicious actors—can access intimate details of people’s lives without their consent? This isn’t just a technical oversight; it’s a glaring vulnerability that could be exploited in countless ways.
The exposure was first discovered by YouTuber and technologist Benn Jordan, who shared his findings with security researcher Jon ‘GainSec’ Gaines. Together, they used Shodan, an IoT search engine, to identify the improperly secured cameras. Jordan recounted his initial reaction: ‘Immediately, we were seeing everything—from playgrounds to parking lots, people Christmas shopping, unloading their stuff into cars. I think the one that affected me most was the playground. You could see unattended kids, and that’s something I want people to know about so they can understand how dangerous this is.’ In a YouTube video, Jordan demonstrated how easily someone could use the exposed footage to identify individuals using open-source tools, highlighting the potential for abuse.
To verify these claims, I drove to Bakersfield and walked in front of two of the cameras while watching myself on the livestream. I also reviewed Flock’s contracts with cities, analyzed company presentations, and geolocated several of the exposed cameras across the U.S. Jordan even filmed himself in front of cameras on the Peachtree Creek Greenway bike path, further confirming the scope of the issue.
Here’s the bigger question: As AI-powered surveillance becomes more pervasive, who is responsible for ensuring these systems are secure? And what safeguards are in place to protect the public from unintended—or worse, malicious—exposure? Flock’s exposed cameras aren’t just a technical failure; they’re a wake-up call about the risks of unchecked surveillance technology. What do you think? Is this a necessary trade-off for public safety, or a step too far into invasive monitoring? Let’s discuss in the comments.
