The physical security industry finds itself at a crucial juncture, with the rapid transition of video surveillance and analytics to the cloud offering both enhanced connectivity and intelligence. However, alongside these advancements, concerns arise regarding mass data collection, profiling, and potential abuse. Verkada, a leading cloud-based provider in the industry, is taking the initiative to chart a path that prioritizes privacy amidst these emerging tensions. By introducing features focused on protecting identities and validating footage authenticity, Verkada aims to establish a privacy-first approach.
Privacy-Focused Features to Protect Identities
In an interview with Verkada founder and CEO Filip Kaliszan, the motivation behind the new privacy and verification features is outlined. Kaliszan emphasizes their mission of protecting people and property in the most privacy-sensitive way possible. He states, “[The feature release] is about that privacy-sensitive way of accomplishing our goal.”
“Our mission is protecting people and property in the most privacy sensitive way possible,” Kaliszan said. “[The feature release] is about that privacy sensitive way of accomplishing our goal.”
The first update focuses on obscuring identities in video feeds. Verkada cameras will gain the ability to automatically blur faces and video streams, akin to augmented reality filters on social media apps. Kaliszan highlights that security guards monitoring feeds do not necessarily require all the intricate details about individuals until an incident occurs. Therefore, making blurring the default path whenever possible is a priority, with the goal of obfuscating identities in most videos.
Video Validation and Authenticity
In addition to facial recognition-based blurring, Verkada plans to implement hashing of captured video to create a tamper-proof digital fingerprint for each video. This feature helps combat concerns surrounding generative AI that can potentially fake or alter footage.
“We can say this video is real. It came out of one of our sensors and we have proof of when it was captured and how, or hey there is no match,” Kaliszan said.
By ensuring video authenticity, Verkada integrates privacy and verification capabilities while also aligning with ethical imperatives. Kaliszan emphasizes that it’s a win-win strategy for Verkada, as they do what they believe is right for society while building customer trust and preference.
However, some critics argue that Verkada’s incremental changes do not go far enough to protect privacy. Merve Hickok, president of the independent nonprofit Center for AI and Digital Policy, suggests that companies like Verkada should adopt a privacy-enhancing approach that avoids collecting unnecessary data in the first place. Hickok believes that even blurred footage can still enable tracking through location data, license plate readers, and heatmapping.
“If you’re doing it where it can be undone — you can undo it later — you’re still collecting that very intrusive information,” Hickok said.
Without stronger regulations, there is a concern that society is heading towards ubiquitous public surveillance. Hickok advocates for legal prohibitions on real-time biometric identification systems in public spaces, akin to those being debated in the European Union.
The Balancing Act and Implications for the Security Industry
Verkada finds itself at the crossroads of different perspectives on ethics and technology. While Kaliszan aims to demonstrate that security can be “privacy-sensitive” through features like blurring, critics argue whether Verkada’s business model can fully align with individual rights.
This dilemma holds significant implications not only for Verkada but also for the broader security industry. As physical security continues its transition to the cloud, companies like Verkada are guiding numerous organizations into new technological territory. The choices made today regarding data practices and defaults will have far-reaching consequences.
According to Hickok, companies need to be cognizant of their decisions and strike a balance between security and mass surveillance:
“We’re way closer to enabling the fully surveyed society than we are from a fully private and protected society,” she said. “So I think we do need to have that security measure, but maybe the takeaway here is the companies just need to be very cogent.”
For Verkada, achieving cogency entails advancing security while avoiding the development of a mass surveillance system. By considering privacy, tying identity locally, and processing data at the edge, Verkada aims to align their approach with societal needs and expectations.
“When all of it comes together, that privacy consideration further increases, right?” Kaliszan said. “And so thinking through how do we maintain privacy, how do we tie identity locally, doing the processing on the edge and not building a mass surveillance system.”
