Keeping humans in the loop: the future of emergency management

[...]

In the context of emergency response, JCS means that emergency managers and AI agents work together, each contributing their unique capabilities to the collective response effort. AI agents, which are computer systems that can independently perform tasks and make decisions, can rapidly process and analyze large quantities of data, providing real-time situational awareness and predictive capabilities. Emergency managers, in turn, bring their expertise in decision-making, critical thinking, and leadership. Achieving this would require the reinvention of roles, with a heightened focus on responsibilities that demand a trusted human presence. Other high-risk professions that have successfully integrated automation while preserving the criticality of human-directed work offer valuable insight. For instance, despite the advanced capabilities of autopilot systems, passengers still anticipate being greeted by a reassuring captain upon boarding their flight.

[...]

While AI integration holds great promise, it also presents critical challenges. One well-known issue is data quality and bias, as AI systems rely on the accuracy and comprehensiveness of their training data. Flawed data can lead to unreliable outputs, emphasizing the need for continuous efforts to ensure high-quality datasets and mitigate biases. By using the JCS approach, humans are more aware of how the entire system operates and, therefore, can better identify weaknesses.

Another challenge is system complexity, which increases the likelihood of unexpected failures. Emergency managers must understand the limitations of AI systems and their potential failure modes while maintaining the ability to perform tasks manually or with alternative methods when necessary. Despite significant progress toward "on-device AI," many of the most robust systems rely on network connectivity to a cloud-based service.

[...]

To fully utilize AI, strategic leaders should actively shape its integration by focusing on two critical areas: workflow redesign and workforce reskilling. Workflow redesign involves a comprehensive analysis of existing processes, breaking them down into their individual components. This deconstruction allows for the identification of bottlenecks, inefficiencies, and tasks that are particularly well-suited for AI augmentation or automation. For instance, AI excels at rapidly processing large datasets, making it ideal for tasks like risk assessment, predictive modeling, and resource allocation. Automating routine tasks, including report generation or data entry, can free up personnel for more strategic and people-facing activities.

[...]