AI-driven lighting systems could cut energy use by over half

A new academic review argues that artificial intelligence (AI) could help cities dim the lights intelligently, cutting both energy consumption and light pollution without compromising visibility or safety.

In their study titled “AI-Driven Adaptive Urban Lighting for Reducing Light Pollution and Energy Consumption in a Multi-Level Perspective,” published in Energies, researchers present a conceptual framework that blends artificial intelligence, Internet of Things sensor systems, and sustainability transition theory to rethink how urban lighting operates.

Energy burden and environmental cost of artificial light

Urban lighting serves essential functions. It supports nighttime mobility, public safety, commercial activity, and social life. It also accounts for a significant share of municipal electricity demand. Many cities continue to rely on centralized control systems that follow fixed schedules and uniform illumination levels, regardless of pedestrian flow, traffic density, or natural light conditions.

Even where LED technology has replaced older lamps, the full efficiency potential often remains untapped. Studies show that LED retrofitting alone can deliver energy savings of 30 to 50 percent. However, substantially larger reductions, frequently exceeding 60 percent, are achievable only when intelligent control systems dynamically adjust brightness in response to real-time data.

The environmental consequences of excessive lighting extend well beyond energy waste. Artificial light at night, commonly referred to as ALAN, is one of the fastest-growing forms of environmental pollution. Satellite data reveal steady increases in urban brightness and illuminated land areas worldwide.

Ecological research has linked excessive nighttime illumination to disrupted circadian rhythms in animals, altered predator–prey relationships, and changes in reproductive and migration patterns. Nocturnal insects, birds, amphibians, and mammals are particularly vulnerable. Plant growth cycles can also shift under prolonged artificial illumination, affecting ecosystem stability and biodiversity.

Human health concerns add another layer of urgency. Chronic exposure to artificial light at night has been associated with sleep disruption, circadian misalignment, and increased risks of metabolic disorders and cardiovascular disease. Melatonin suppression and hormonal imbalances are frequently cited mechanisms. Despite mounting evidence, lighting policy has often prioritized visibility and compliance standards over ecological and health considerations.

The authors highlight that research on urban lighting has been fragmented across disciplines. Engineering studies focus on energy metrics and control algorithms, while environmental research documents harm without fully integrating technological mitigation strategies. This gap, they argue, has slowed systemic reform.

AI and IoT as catalysts for urban lighting transformation

These systems integrate sensor-equipped luminaires, IoT networks, and machine learning algorithms to modulate illumination in real time. Instead of operating on fixed schedules, adaptive systems adjust brightness according to traffic levels, pedestrian activity, weather conditions, and ambient light.

Machine learning models can predict activity patterns based on historical data, enabling proactive rather than reactive lighting control. Computer vision technologies can detect movement, while real-time dimming algorithms optimize energy use without sacrificing safety.

Pilot deployments and experimental programs suggest that adaptive systems can significantly reduce electricity consumption while maintaining compliance with safety standards. Moreover, when integrated with renewable energy sources and smart grid infrastructure, lighting systems can function as flexible assets within broader urban energy networks.

However, the authors caution that technology alone will not drive transformation. They analyze adaptive lighting through the Multi-Level Perspective, a widely used framework in sustainability transition research. The MLP distinguishes among three analytical levels: niche innovations, dominant regimes, and landscape pressures.

At the niche level, AI-driven lighting systems represent experimental innovations developed in protected contexts such as smart city pilots or sustainability-oriented districts. These niches allow experimentation with reduced institutional pressure, fostering learning and iterative design.

However, scaling beyond niche status requires alignment with regime-level institutions. Municipal procurement processes, regulatory norms, and operational routines often favor incremental improvements rather than disruptive change. Risk-averse decision-making, sunk infrastructure costs, and standardized safety rules reinforce stability within the existing lighting regime.

The authors emphasize that successful adoption depends on both technological maturity and social legitimacy. Participatory governance, municipal incentives, and public awareness campaigns are critical for building trust and acceptance. Adaptive lighting systems must demonstrate not only efficiency gains but also reliability, safety, and fairness.

Governance, policy, and the multi-level transition framework

The regime level of urban lighting includes entrenched technologies, regulations, professional standards, and institutional arrangements. Historically, lighting systems were designed for uniform brightness and centralized control. These design principles remain embedded in procurement contracts and safety regulations, limiting flexibility.

Municipal authorities, utility providers, urban planners, and lighting manufacturers collectively shape this regime. Cognitive and cultural factors, including perceptions of safety and tolerance for lower brightness levels, further stabilize the status quo.

The study argues that regime inertia is a major barrier to transformative change. Institutional lock-in and compliance-driven governance can constrain innovation even when technical solutions are available. Without regulatory reform and adaptive policy instruments, AI-driven systems may remain isolated experiments.

At the landscape level, broader societal forces exert pressure on the regime. International climate agreements, national energy efficiency targets, rising electricity prices, biodiversity loss, and growing public awareness of light pollution create conditions favorable to reform.

Smart city agendas and advances in AI and IoT technologies further destabilize existing structures. Disruptive events such as energy crises or urban redevelopment initiatives can accelerate adoption by exposing inefficiencies in traditional systems.

The authors propose three guiding research questions: how path dependencies influence adoption, how regulations shape energy and light pollution outcomes, and how technological advances can support policy-driven transitions. Their analysis remains conceptual, focusing on interactions rather than empirical case studies.

Business model innovation emerges as a crucial enabler. Public–private partnerships, energy performance contracting, and service-based lighting models can distribute financial risk and encourage adoption. Integrating adaptive lighting into smart grids and renewable energy systems strengthens the economic rationale.

The study also acknowledges limitations of the Multi-Level Perspective. While MLP provides a structured lens for analyzing socio-technical transitions, it may underrepresent agency, political contestation, and power dynamics. Stakeholders such as utility companies or contractors may resist innovation due to vested interests.

Consequently, the authors call for complementary frameworks from institutional theory and political economy to better capture governance complexities. They stress that sustainability transitions are rarely linear or orderly. Power struggles and institutional resistance can shape trajectories in unpredictable ways.

Equity considerations are also vital. Transition strategies must address fair distribution of costs and benefits, equitable access to lighting, and community engagement. Performance indicators should include not only energy savings but also reductions in skyglow, mitigation of ecological disruption, and improvements in public perception.