How to Navigate the Risks and Benefits of AI-Based Video Surveillance in Buildings

Building owners and facilities managers fight an uphill battle to retain tenants, protect their properties, and achieve a favorable profit margin. Can the promise of artificial intelligence (AI) help them get the upper hand?

Recent years of industry turbulence have reached a fever pitch for facilities professionals:

• Insurance costs have skyrocketed, in part, due to property damage from natural disasters and vandalism.
• Evictions across the U.S. are rising as pandemic-era support systems wind down.
• Physical and cybersecurity risks remain high, especially as smart building technology and connected devices become more widespread.

It is not easy to get ahead of these issues and increase property control without harming the tenant experience. It requires the right mix of human expertise and technology.

As you consider how to improve your building control, let’s demystify a few things around AI and explore five steps to help you decide if AI is right for your building. 

Understanding the Benefits and Risks of AI for Property Management

AI can process massive amounts of data, recognize patterns, and present findings or recommendations based on statistical probabilities from its input. AI can do a lot to relieve your team from repetitive tasks or augment what they can accomplish.

The risk, however, is that teams may defer key decision-making to AI. Or, they “set it and forget it,” allowing potential problems to spread until someone catches the error. Remember: A human must always stay in the loop.

The best AI applications make human jobs easier and free them up to do what they do best, which is why video surveillance is a key area to consider. In a fully controlled environment, AI-powered video solutions can provide building owners with insights like:

• When an individual who is banned from the property enters the premises 
• Typical property activity levels by day and time, including in communal areas like playgrounds or parking lots
• Whether there is suspicious activity on the property, like unattended packages or potential illegal activity

A human being paid to stare at a video monitor is not effective. By nature, humans are inept at finding specific items that appear infrequently (a concept called sporadic visual search). AI computer vision can automate evidence analysis to flag items or footage that a human should review. If a human can detect it, AI can be trained to detect it, too!

Of course, tenants may have understandable concerns regarding how you employ AI:

• Anonymity: Can they be identified from the footage? Is there facial recognition?
• Data: Is their activity or behavior being sold or used in some way?
• Privacy: Are they being watched at all times?

Further, teams invite potential cybersecurity risks if they fail to fortify their systems and protect their customer data. This is why it’s essential to craft an AI strategy tailored for your implementation and proactively address potential tenant concerns. 

5 Steps to Assess Where AI Fits in Your Video Surveillance Strategy 

Each building environment carries unique risks and tenant expectations. To decide where AI can fit to best support your goals, follow these five steps:

1. Audit Existing Surveillance Capabilities

First, inventory all of your existing surveillance and building security solutions. 

• What does each solution offer? 
• Are other features available from that provider? 
• Does the solution plug into your ecosystem, or does it operate independently?

Create a spreadsheet and list all solutions, primary use cases, and shortcomings or challenges with the solution. If appropriate, list each solution’s cost to see where consolidation could streamline your resources.

Speak with your security team and property staff to understand how they use each solution (if at all) and what opportunities they see to strengthen security. 

2. Assess Capability Gaps or Overlaps

Review your inventory to see where you’re already using AI. Can you expand that use case to more areas on your property? Do you need a complementary solution to work alongside your existing stack? 

Solution vendors frequently add new capabilities or features, so your team may not be using its existing toolkit to the fullest potential—or overpaying for redundant solutions. 

This is also an appropriate time to decide whether you want on-premise or cloud native video solutions to run your AI video analytics. On-premise solutions typically require significant updates to introduce AI capabilities. However, home built solutions offer the greatest customizability and control (assuming you have the engineering talent). Cloud native solutions can more easily scale to your needs and deploy the latest AI models, usually with a lower total cost of ownership. 

3. Determine Your AI Strategy

With an understanding of your existing capabilities, your team can now determine what you want AI to accomplish. Ask yourself the following questions:

• What notable challenges does your team currently face? 
• What gaps exist in your existing security strategy and toolkit?
• Do specific processes take more human time than you can afford? 
• What is a high impact challenge you can solve to drive a noticeable improvement in your building control? 

Identify one or two areas where you can safely introduce AI for relatively low effort to begin testing its use cases and capabilities. Especially if you are new to using AI-powered solutions, it’s best to start with one or two controlled use cases before expanding your scope. 

4. Establish Data Governance and Control Procedures

What data is collected from your employees, tenants or the public to train the AI system for your use case(s)? Where does this data come from, who can access it and how often is it updated or deleted?

Your team needs to understand what data is used and identify any potential risks about how it is collected or how it could be accessed by nefarious players. A key part of this process, too, is to assess how an AI model is trained to ensure the resulting system has as little bias as possible. 

5. Measure and Refine As You Go

New risks and opportunities emerge each day as our communities change and technologies become smarter and more powerful. 

Start a regular audit cycle with key stakeholders to revisit your strategy and explore your primary risks, any on the horizon, and how your technology is addressing these challenges.

Increasing Control of Your Building or Facility

Technology should help speed things up, make things cheaper, or improve accuracy and consistency. If there isn't a targeted goal to achieve one of those outcomes, then the technology solution shouldn't be implemented.

AI can be incredibly powerful for extending your video surveillance and augmenting your existing security investments, but you must be ready to mitigate its risks. The best solution will blend human expertise with the power of AI to help you cover more ground than ever and provide tenants with a superb experience.