Artificial intelligence (AI) is transforming the landscape of surveillance and monitoring, delivering powerful tools for security while posing complex operational and ethical challenges. The Monitoring Association (TMA) brought together a panel of leading experts about AI in monitoring centers at its November 2024 OPSTech meeting — Actuate Co-Founder and CEO Sonny Tai; Dice Corporation Co-President Avi Lupo; and Watchful Founder Joshua Parsons shared their perspectives on the current state and future potential of AI in the field. Here are five key insights and takeaways from the experts aimed at setting your monitoring center up for successful adoption of AI technology.

1. Adopt Generative AI Thoughtfully

Generative AI offers immense potential to automate tasks, enhance decision-making, and reduce operational workloads, but its integration must be carefully managed. Companies should focus on:

• Implementing guardrails — Establish robust validation systems to ensure AI outputs are accurate and contextually appropriate. Operators should not become overly reliant on AI-generated results, which could lead to errors if unchecked.
• Maintaining human oversight — AI should act as a support tool rather than a replacement for human expertise. This balance will ensure accountability and improve trust in AI-driven decisions.
• Piloting and iteration — Begin with pilot programs that test generative AI’s impact in controlled environments. Use these to refine the system, address unforeseen issues, and optimize workflows.

2. Invest in Training & Development

The panel emphasized the importance of comprehensive education for both internal teams and clients to bridge the gap between technological possibilities and operational realities. For example:

• Internal training — Equip design and sales teams with the knowledge to accurately represent AI’s capabilities and limitations. This prevents overpromising and fosters better alignment with customer needs.
• Customer education — Help end users understand how AI systems work, their limitations, and the benefits they can realistically expect. Clear communication reduces dissatisfaction and builds long-term trust.

3. Start Small & Scale Gradually

Rather than committing significant resources to large-scale deployments, businesses should take a phased approach to AI integration with:

• Small pilots — Start with limited-scale implementations to test AI solutions in real-world conditions. This allows teams to gather insights, refine operations, and correct inefficiencies without incurring massive risks.
• Iterative scaling — Expand operations incrementally, using feedback from each phase to improve the technology and align it with business needs. This strategy minimizes disruption while maximizing learning opportunities.
• Lean infrastructure investments — Avoid large upfront infrastructure commitments until the technology’s viability and scalability have been demonstrated.

4. Focus on Practicality & Immediate Needs

The panelists highlighted a common industry pitfall: overhyping future capabilities while neglecting solutions that work effectively today and advised monitoring centers leadership to:

• Deliver reliable basics — Focus on achieving high accuracy in core functionalities, such as human presence detection or vehicle monitoring. Addressing these foundational needs at a 99 percent plus reliability rate creates value for customers and sets the stage for more advanced applications.
• Avoid overpromising — Recognize the technological limitations of advanced analytics and behavioral predictions. Provide transparent expectations to clients about what AI can realistically achieve in the short term.
• Customize solutions — Tailor AI implementations to the specific needs of the customer, such as focusing on false-positive reduction or optimizing workflows for particular industries.

5. Address Ethical & Privacy Concerns

As AI systems become more prevalent in surveillance, companies must proactively address ethical and legal challenges to maintain trust and compliance, including:

• Bias mitigation — Implement rigorous checks and diverse training datasets to reduce racial or demographic biases in AI models. Proactively test outputs for fairness and retrain models as necessary.
• Privacy protections — Clearly define data ownership and ensure compliance with legal frameworks like GDPR and BIPA. Offer flexible storage options, such as local or customer-controlled systems, to address privacy concerns.
• Transparent policies — Be upfront about how customer data is used, including whether it is leveraged for AI training. Ensure customer consent for any secondary uses of their data.
• Industry collaboration — Advocate for standardized guidelines in AI usage for surveillance to build trust and consistency across the industry.

Real-World Successes & Emerging Business Models

The panel shared concrete examples of AI success. Parsons described a monitoring station in Houston that improved its efficiency by combining cloud AI with generative AI, reducing its operator-to-camera ratio from 1:450 to 1:1800. Tai highlighted customer demands for reliable basic detection systems, which form the foundation for more advanced analytics in the future.

The discussion also touched on evolving business models, with panelists debating the merits of per-camera pricing versus resource-based billing. The latter approach, they argued, could drive fairness and efficiency by aligning costs with actual usage.

Looking ahead, the panelists emphasized the need for businesses to adopt open-minded approaches to AI and automation. Lupo advised against making large infrastructure investments without understanding rapidly evolving technologies. Tai advocated for a “startup within a startup” model, encouraging companies to start small and scale gradually while learning from hands-on implementation.

Exciting trends in AI applications, such as shoplifting detection based on body behavior analysis, were discussed, but the panelists agreed these technologies are still in their infancy. 

This panel presentation underscored the transformative potential of AI in surveillance and monitoring, balanced by the practical and ethical challenges of deployment. As the industry continues to evolve, a measured, customer-focused approach will be crucial to unlocking AI’s full promise.

Key Recommendations for Industry Players

• Adopt Generative AI Thoughtfully: Ensure guardrails are in place to prevent overreliance by operators.
• Invest in Training: Equip teams with the knowledge to design and sell AI solutions effectively.
• Start Small: Pilot projects to refine strategies and build scalable solutions.
• Focus on Practicality: Prioritize technologies that deliver reliable results today over speculative future capabilities.
• Address Ethical Concerns: Proactively mitigate biases in AI models to ensure fair outcomes.