printer friendly version

From the editor's desk: We’ve only just begun

Andrew Seldon, Editor.

The surveillance market has expanded far beyond the analogue days of just recording and/or monitoring screens. The capabilities of surveillance technology today extend to black screen monitoring with video analytics enhanced by artificial intelligence, and with analytics and AI taking CCTV beyond the security realm.

For companies and cities with well-run and maintained surveillance installations, the possibilities are endless, from traffic management to a rapidly advancing field of individual and crowd behavioural analysis. Behavioural AI still has a way to go, but the potential is enormous, as are the risks.

Unlike humans, AI decides and cannot re-evaluate that decision on the fly without prompting. As a simple example, AI might see two people hugging as a violent act and raise an alert, whereas humans will be able to understand the context of the scene and ignore it.

Those of us who have been fortunate to see some of Craig Donald’s presentations on control room operator training can easily see the benefits of accurate behavioural analysis that alerts operators to a potential incident before something happens. These events don’t always require the latest high-resolution video, but rather the understanding of why a person or people are standing and behaving in certain ways.

The concept of context is one of the fields being studied and worked on with gusto, and AI’s ability to ‘understand’ a scene may soon become a reality. It will take a few years before the technology is publicly available, but we will get there. Well-trained operators are not at risk of being replaced by AI just yet.

Then we get agentic AI, which is also in its infancy, but which could and will change the way the world works. If you look at the so-called chatbots on many websites, their efficiency is even poorer than Johannesburg’s municipality, so there is a long way to go. Creating an AI to fill craters in the road (they are too big to be called potholes), would be a brilliant invention.

The advancing technological capabilities not only provide more functionality and services to the end user, but also more complexity and require better planning. The late Rob Anderson always used to say you need to give every camera a job description, and this is more important than ever. If the end user can’t describe exactly what a camera is supposed to do, there is a good chance that they will spend more on expensive technology that is never fully leveraged. The opposite may also happen, with the user getting a good deal on technology that doesn’t do the job they thought it would, or the job they decided (after the installation) that it should do.

Technology also requires cooperation between different departments. If access control can tell event managers how many people are onsite, where they are, and what they are doing (via integrated surveillance) they benefit – of course, people counting and whereabouts is also critical for emergency evacuation.

That said, I hope you enjoy the SMART Surveillance Handbook 2025, and as always, any comments and criticisms are welcome.

Share this article:

Further reading: