Beyond standard pressurisation systems, AI is enabling smoke control that reacts to a fire's location and spread. We review the technology and its implications for design.. Next Gen Smoke Control: The Rise of AI Powered Dynamic Systems The UK fire safety landscape is on the cusp of a significant transformation, as artificial intelligence (AI) begins to move beyond theoretical discussions and into practical applications for smoke control systems. No longer confined to static, pre programmed responses, next generation AI powered dynamic systems promise a revolutionary leap in occupant safety, offering real time adaptability to the unpredictable nature of fire. This innovative approach, which leverages machine learning and sophisticated algorithms, is poised to redefine how fire engineers design and implement smoke management strategies in complex, high risk buildings, moving beyond the limitations of traditional pressurisation and natural ventilation systems to offer a truly intelligent response to fire incidents. Background For decades, smoke control in the UK has relied primarily on established principles and prescriptive guidance. Approved Document B (ADB) provides foundational requirements, while standards like BS 9991 and BS 9999 offer detailed methodologies for smoke ventilation and management in various building types. These frameworks typically advocate for systems designed to achieve specific performance criteria under predefined fire scenarios. Common solutions include mechanical smoke extract systems, natural smoke ventilation, and pressurisation systems, particularly in protected escape routes and stairwells. The Regulatory Reform (Fire Safety) Order 2005 (RRO 2005) places the onus on responsible persons to ensure adequate fire safety provisions, including effective smoke control, are in place and maintained. However, a fundamental limitation of these traditional approaches is their static nature. Designs are based on a 'design fire' – a theoretical fire size and location – and the system operates according to pre set parameters. In a real world incident, a fire's characteristics, growth rate, and precise location can deviate significantly from these assumptions. Factors such as compartmentation breaches, unexpected airflow paths, or the presence of unusual fuel loads can render a static system less effective than intended. This inherent inflexibility means that while traditional systems are vital, they may not always provide the optimal response to a dynamic and evolving threat. The tragic events of recent years have underscored the critical need for more robust and adaptable fire safety measures, driving a renewed focus on innovative solutions. Key Developments The emergence of AI in smoke control addresses this critical gap by introducing dynamism and intelligence. At its core, an AI powered dynamic smoke control system integrates a network of sensors (smoke detectors, heat detectors, CO/CO2 sensors, airflow sensors, CCTV with AI powered fire detection) with a central processing unit running sophisticated algorithms. This system continuously monitors the building environment. When a fire is detected, the AI doesn't simply activate a pre set sequence; instead, it performs real time analysis: 1. Precise Fire Localisation and Characterisation: Using data fusion from multiple sensors, the AI can pinpoint the fire's exact location, estimate its size, and predict its likely spread path. This goes beyond simple zone activation, offering granular detail. 2. Dynamic System Optimisation: Based on the real time fire data, the AI intelligently adjusts the smoke control system's operation. This could involve: Variable Fan Speeds: Instead of fixed extract rates, fans can be modulated to precisely control pressure differentials and extract volumes, optimising smoke clearance and preventing smoke migration into escape routes. Targeted Damper Control: Individual dampers can be opened or closed strategically to create specific airflow paths, isolating smoke to the fire floor or compartment while maintaining tenable conditions in adjacent areas. Integration with Other Building Systems: The AI can communicate with and control other building systems, such as HVAC, lifts, access control, and even emergency lighting, to coordinate a holistic response. For example, it might direct lifts to safe zones, unlock specific doors for evacuation, or adjust lighting to guide occupants. Predictive Modelling: Advanced AI can use predictive algorithms to forecast smoke movement based on current conditions and building geometry, allowing for proactive adjustments to mitigate future smoke spread. 3. Human Machine Interface: Modern systems provide intuitive dashboards for fire command teams, offering real time visualisations of smoke movement, system status, and AI generated recommendations, enabling informed decision making. 4. Learning and Adaptation: Over time, AI systems can "learn" from fire drills and real incidents, refining