A transformation driven by intelligence
Kasia Hanson, Global Sr. Director, Physical and Cybersecurity Ecosystems, Intel Corporation – in partnership with Omdia’s Niall Jenkins and Scott Foley – explores AI outcomes in the security market and beyond.
It’s undeniable. The security industry is transforming. This evolution has been sparked by the momentum of AI, which is amplifying the power of data.
Through data analysis and inference, AI can dramatically increase insights and efficiency, opportunities and outcomes.
And, with AI enabling the intelligent analysis of vast amounts of data, it’s also leading the shift towards edge computing.
The combination of AI and edge computing (Edge AI) brings AI’s insights and decision-making capabilities closer to data sources, such as cameras and sensors.
Edge AI can make a positive impact through immediate inference as well as the ongoing customisation of pre-trained models based on an organisation’s unique data.
By utilising machine learning models on local or edge devices, real time decisions are possible without sending data to and from the cloud.
Instead, Edge AI can process and analyse data where most of it – 85% estimated by Canalys – is already located, on-prem rather than in the cloud.
This doesn’t mean the cloud doesn’t have a role in providing AI-enabled analytics for the industry.
In fact, a hybrid AI approach combines the benefits of the edge with the deeper context of the cloud. Using this method can facilitate automated decision-making and self-healing systems.
Of course, no examination of security’s transformation would be complete without mentioning AI’s impact on video, which has become the principal edge AI workload.
This is due to the advanced AI and intelligence employed through both video cameras and video analytics.
With one billion cameras worldwide, this trend is likely to continue – allowing cameras to be used in new ways…
The journey towards holistic AI-enabled security
The AI-fuelled transformation of the security industry is indeed a journey, rather than a destination.
AI adoption and the shift towards security and business intelligence at the edge will vary across time and companies.
Yet the main stages of this journey towards a holistic, insight-driven approach to security are generally the same, regardless of a business’s circumstances.
The first two stages represent the digitisation phase of the journey, while the final three stages represent the AI insights phase – when the promise of AI is realised.
To help identify where you are in this maturation process, we’ve provided a synopsis of each stage.
This information can help you determine the type of expertise and solutions you need today – and in preparation for tomorrow.
Phase I: Digitisation
Physical security – this stage is focused on the basic protection of people, places and assets. It includes traditional security measures, such as access control systems and cameras, to establish a defence perimeter, deterring and detecting unauthorised access. These systems are often analog, moving to IP or both – and they may require labour-intensive manual monitoring.
Edge-optimised security – organisations enter this stage as they begin adopting security architecture, software and IP-enabled devices that are optimised for the edge. These solutions enable connected security, enhance access control, optimise operations and real-time monitoring. By processing data closer to its source, this security also reduces latency and increases responsiveness.
Phase II: AI insights
Camera as sensor – at this stage, video cameras evolve from passive recording devices to active sensors capable of analysing and interpreting data. Working as connected devices that leverage the power of AI, cameras as sensors can identify and predict threats – evolving security from reactive to proactive. These cameras may be deployed in traditional physical security locations as well as integrated with robotics or UAVs. And, by acting as sensors, cameras can also offer intelligence beyond security, providing a range of business outcomes.
AI everywhere – with the proper infrastructure and upgraded devices in place, businesses can enter the stage where AI is integrated across all aspects of the security landscape. The prevalence of AI can enable sophisticated capabilities, such as predictive analytics and automated responses, while also helping to align security with business operations. Solutions can be deployed to provide contextual and behavioral identification as well as greater intelligence and enhanced outcomes. This applies to a range of verticals, from manufacturing and retail to education and healthcare.
Holistic security – at this stage, physical and cybersecurity converge into a unified approach. Holistic security solutions are designed to address the expanded attack surface created by interconnected devices and systems. In fact, device security is a key use case for these solutions. Others include managed detection and response, OT security, data centre protection and supply chain transparency. These solutions also leverage the latest principles, such as zero trust, confidential computing and endpoint protection, helping to secure assets.
To help security practitioners and integrators on this journey, Intel is partnering with ecosystem partners to deliver AI workshops, advisory services and AI customer tools to help the industry navigate purposeful adoption through the security lifecycle.
Intel and Omdia (Authors – Niall Jenkins and Scott Foley) partnered to further explore how AI is creating new outcomes in security and changing the way businesses and organisations operate across the consumer, automotive, industrial and government markets.
AI is a universal business opportunity for the tech industry.
Omdia and Intel: AI Outcomes Report
This report aims to provide a background on the foundational security infrastructure on which the AI growth journey can be built.
It also looks to analyse the key technology trends and vertical roadmaps that can be used by integrators and other security professionals to deliver successful AI outcomes.
In the video security market, cameras are the primary source of video images for AI and computer vision.
However, the camera and traditional security use cases are only the start of the business opportunity.
Outcomes for marketing, operations, sales, supply chain management and people management functions can all leverage AI combined with existing video security infrastructure.
This is a critical opportunity for integrators and other participants in the video security industry.
Moving beyond security
AI is now prevalent across many traditional physical solutions. This includes perimeter protection and virtual tripwire, facial authentication, object detection, tracking and behaviour recognition.
The next step in the AI journey will be to move beyond traditional security applications.
Leveraging the installed base of network security cameras, the industry has an opportunity to take the lead in this already emerging AI market evolution.
Marketing, business intelligence, operations and customer experience solutions can all be built on the video camera-installed base.
Additional sensors, such as AI-based sound analytics, can also provide inputs, whether that is detection of a gunshot or of an unusual noise in a manufacturing process.
In some cases, additional edge AI compute will be required to support these use cases; dedicated AI appliances or new servers can typically be integrated without impacting infrastructure.
These steps do not need to be taken in one leap either. Many algorithms used in business intelligence analytics have components similar to those used in security applications.
Service providers can learn how to apply new AI solutions in a phased approach, meeting the needs of new roles within existing clients, delivering similar algorithms in new applications and building a position in the operational processes of their customers.
As network cameras become more affordable and the benefits become widely known, the installed base is forecast to increase at double digit growth rates.
The transition from analog cameras with the associated replacement opportunity is also driving growth.
Many analytics used in physical security started development in high end, enterprise markets. Often, this meant a mission-critical system where budgets were readily available to support development costs.
People counting analytics, for example, have been used in crowd management for some time.
Now, they are deployed in more cost-sensitive verticals such as retail and commercial.
The combination of using existing video infrastructure, lower software development costs (the expense of initial development has already been incurred), less expensive compute power, better trained integrators and the economies of scale of selling to larger markets means overall cost of analytics is lower.
This has resulted in traditionally high end analytics being used in operational and business intelligence.
Market trends
Computer vision trends in other industries have the potential to affect security professionals.
GenAI advancements could make data training easier to produce and use.
Segments that share computer vision needs such as smart cities and retail are investing heavily. Outcomes will evolve quickly.
Some examples include advanced neural networks and 3D imaging/LiDAR in autonomous driving, edge computing and facial identification in smart city infrastructure, motion detection and tracking in sports analytics and gesture recognition in retail and gaming.
These evolutions could make keeping pace with the tech more difficult.
However, security professionals that keep pace, expand or specialise in these developments will find themselves as thought leaders, ahead of competition.
Similarly, professionals will increasingly need a foundational understanding of AI principles, machine learning algorithms and their applications.
This requirement could extend to an ability to manage large volumes of data and the skills to extract insights from or beyond current analytics offerings.
Cybersecurity is another area of focus. Ensuring that physical security and cybersecurity requirements are considered together is important because one cannot exist without the other.
This also requires skilled people that understand the role of cybersecurity in the overall security of an AI solution. As cyber-threats continue to evolve, this role will be increasingly important.
Finally, the cost and complexity of AI outcomes is still high.
Enterprises may not find this daunting, but many smaller organisations will continue to use less expensive solutions that have a range of installers and can be maintained efficiently.
Legacy systems can be difficult to displace in the industry. Consequently, the benefits and opportunities inherent in AI must be promoted and justified.
At this point, the market is still relatively young.
Manufacturers, integrators and third party institutions have an opportunity to display leadership and provide a mentoring role.
The rest of the industry will have to invest in training or expand areas to accommodate.
Showing how AI can be applied across different end user industries will form part of the growth story.
To read more of the report, including the vertical approaches, please click here.