DURING every site visit SEN has undertaken this year, end users have expressed concerns felt by their organisations towards video analytics and particularly towards face recognition.
It’s a trend that could strip the electronic security industry’s future of its operational crown jewels. In recent conversations, switched-on senior managers have variously and categorically stated: ‘We won’t be going there’ or ‘our clients have requested we don’t use those features’ or ‘the public has no appetite for those sorts of things’ or ‘we might be able to look at that the future’ and even ‘management doesn’t want to take the risk’. Meanwhile, end users who are interested in analytics are also aware of the issues other security managers are facing and they are all talking about it.
Listening to these sentiments from 4 different end users last month, I realised the security industry hasn’t lost control of the narrative around the use of analytics, it never engaged with the narrative in the first place. From the point of view of operational security teams, the message should be that carefully and transparently applied video and data analytics are finally allowing electronic security and automation solutions to operate at close to their full potential. Failure to get this message out is likely to mean administrators and senior managers will fail to properly weigh the advantages of video analytics during their deliberations – fears of negative press will outweigh operational benefits.
The loss to the industry isn’t just that video analytics of all kinds deliver operational torque and serious gains in efficiency and cost reduction, it’s that many future features at all layers of our product stacks are going to depend on leveraging a technology that may no longer have the trust and support of users. In all my time writing about security technology the industry has never faced an issue quite as thorny as this one and it’s difficult to see us digging our way out of it entirely.
Something that makes the task more difficult is that end users and consumers don’t fully comprehend the capabilities and the limitations of video analytics. Even when empowered by deep learning, they are not omniscient. Furthermore, they inform human response, they are not intrinsically reactive. For many security operations, analytics is most worthwhile when searching for past events, or creating situational awareness in integrated solutions. Another key issue is that security technology is being used outside Australia in ways that concern the public here. For instance, while public surveillance solutions are governed by strict protocols in Australia, this is not the case everywhere.
There’s been no specific action against technologies like face recognition, gait detection and broader analytics at a higher government level in Australia, or in countries like the U.S. and United Kingdom. But the tension is there and it’s growing. If the industry – suppliers and users – can’t move in a responsible and coherent way to allay fears, then its likely the future of video analytics is going to be taken out of our hands.
Something that’s possible is we’ll see use restrictions on use in big corporate and government, while in lower end applications analytics, including face recognition, will be freely available. At one level this is a good thing but at another it would deprive serious security operations of the technologies they need to defend their sites. User self-restriction is another possible answer and this could be managed by the creation of uniform procedures around the management of sensitive data in some applications.
It may also be possible to hold some analytic technologies at arm’s length – by using face recognition or fingerprint matching software contained in users’ smart devices, for example. Even the most socially aware tech users in academia are comfortable using biometrics when these are contained on their own smart devices rather than held in reader databases or cards issued to them by a third party. Security people know this bias towards unsecure smart devices makes no sense, but it’s the way things are.
Exactly how the electronic security industry is going to manage these nebulous issues remains to be seen but an open conversation needs to start between end users, suppliers, integrators and consultants and the sooner it starts the better. The thought we might lose a technology that promises to make sense of our data inputs in real time is not a happy one. But if we take the comments of serious end users across multiple verticals at face value, the risk is real.