AI – A Security Operations Perspective
In this Q&A, John Adams speaks with Ken Francis of Eagle Eye Networks about operational perspectives of artificial intelligence, touching on key functionalities, planning, integration and the importance of the API economy to analytics solutions of the future.
JA: How important is it when planning for artificial intelligence to ensure you maintain an operational perspective and don’t get too swayed by the ‘cool’ factor?
KF. It’s extremely important for end users to maintain an operational perspective and share this with their system integrator and video surveillance vendor. This collaboration will enable the integrator and vendor to provide the expertise and appropriate technology to meet the end user’s specific operational needs. Then, looking purposely at ‘cool’ and how it can be applied to today’s and tomorrow’s challenges, ensures the right balance of operational success and future-proofing a system to grow and last.
JA: There’s an understandable tendency for consultants to cover off AI in specifications in a general way – do you think it would be more helpful if required AI functionality was outlined specifically from the start?
KF: Absolutely. When there are too many cooks in the kitchen, the dish may end up overly spicy, or otherwise negatively affected. Similarly, without clearly specifying the desired AI from the beginning, you could end up with a kludgy fusion that, in many cases, is difficult to use. The specification of AI at the front end enables all tech companies in the stack to effectively contribute around a consistent AI selection and help the end user achieve a fully integrated system that meets its specific needs.
JA: What’s the low hanging fruit when it comes to AI – what AI functionalities offer security teams the most for the least, in your opinion?
KF: The most effective AI is functionality that extends over the greatest group of users. The security analytics of line crossing, advanced motion, and people counting, are proven and widely applicable today. The next generation of widely applicable analytics will come from recognition of people, things, and events.
This next wave of AI is becoming more commonplace, due to high demand from end users and greater supply from companies now offering it. Specifically, the need for recognition-oriented AI is growing in critical infrastructure, citywide surveillance, and large campuses (both corporate or university), primarily to reduce manpower requirements and accelerate incident resolution.
JA: Which are the AI areas with the greatest potential in the future – what functionalities should installers and integrators definitely be covering off?
KF: Our recent study showing the most often-used analytics in cloud video surveillance clearly indicates that physical security remains the priority, but we’re starting to see analytics and AI for more than just security. People are rapidly adopting it to help run the business. Customer service and support appear to be the leading areas of interest, but anywhere a business can glean information to improve operational efficiency and effectiveness, reduce maintenance costs, and lower risk are all areas to cover off.
JA: An issue with planning for AI is breadth of lateral application – it very quickly stops being about security when skipping between functions – for instance, AI that allows CCTV cameras to establish recent contacts, check security patrol schedules, identify problem gamblers, etc. How important is it that these sorts of lateral applications are considered in advance by security integrators and their technology partners, and how can security integrators best promote lateral applications that diverge so widely from the security function?
KF: It’s absolutely critical that real problem-solving solutions be designed in advance of installation. The value-added security integrator is key to the design process. He/she brings the know-how to the table to help create a long-term strategy for connecting and managing the various security components, in the most cost-effective way, to achieve optimal system functionality.
JA: A quirk of AI is that it is making its appearance at multiple points in the product stack – with optical and thermal CCTV cameras and their viewers, in touchless disarm alarm panels, in access control reader modules of multiple types, within NVRs, within apps and within management solutions. Is it possible to integrate all these disparate versions of AI, or do end users need to decide well in advance how they want AI to apply to their overall solutions in order to avoid creating networks of high touch points?
KF: This is a great question. Early selection of AI in the video system enables the video system tech company to provide an application programming interface (API) for the select AI, enabling all tech stack participants to share in the AI-generated events. The sharing of a common API is the best method for long-term stability and support. The API should be thought of as the central component within an economy.
JA: How important is it that AI solutions can integrate with security and related subsystems at one level, and with management systems at another level?
KF: AI has moved from being about technology to becoming a critical part of business strategy. Securely connecting mission-critical business applications brings the ability for a business to collect and distil data to applicable business intelligence. The businesses that lead the way are those that can bring distributed and fragmented systems and subsystems together to cross-reference and quickly find information across a multitude of mediums that can lead to better decision making, services, products, etc. In video surveillance, this is happening most successfully in open, true cloud platforms that have the ability to bring partners and technology together to make the world safer and businesses more efficient and effective.
JA: In your experience, how eager are end users to apply AI and what sorts of AI functions are of most interest to them?
KF: Over the past few years we’ve been impressed with the way large businesses and enterprise customers have quickly understood the value of AI and its application to security and business challenges. The small- to medium-sized business owner might have understood the potential of AI, but for many it seemed out of reach because of the infrastructure costs and challenges of legacy security systems. That is changing.
The ubiquity and scalability of our true cloud platform democratizes and makes AI more affordable and accessible for virtually every type and size of business. The conversation has changed from ‘can we do that?’ to ‘what is possible?’ Initially the discussion is typically centred around improving security, but as more people understand what’s possible, it becomes about business insight and optimization.
JA: If there was one vital attitude to bring to the AI empowered future, what would it be?
KF: Open architecture/open platforms are terms commonly used across tech industries. API economy is the next generation of this terminology, and a well-known movement in tech. The API economy will enable organizations to find innovative new ways to extend and complement their services and create value for customers. In the big scheme of things, the API economy will help the video surveillance industry make businesses smarter and the world a safer place.