Abstract
This study investigates the potential for artificial intelligence (AI)-based systems to improve
air quality in smart cities towards sustainable development. To achieve this aim, the study
develops a systematic literature review (SLR)-backed conceptual framework, empirically
evaluates the perception of Ajman residents based on this framework and develops a novel
AI-based solution capable of monitoring and controlling air quality for smart cities based on
data from seven selected environmental stations across the Ajman Emirate. A positivist
worldview is considered in a multi-method quantitative methodological orientation
involving a survey and AI experiment to test the SLR-backed framework. The study
considers a case study of seven neighbourhoods within the Ajman Emirate; these include
Ajman Industrial Area 1 (Al Sinaiyah 1), Al Jerf / Jurf Industrial Area 1, Masfout (All
suburbs), Al Bustan (Liwarah 1), Manama (All suburbs), Mushairif and Hamadiyah. These
neighbourhoods harbour approximately 20% of the Ajman population to the tune of 100,918
residents. Whilst a sample of 383 residents is considered generalisable based on Slovin’s
formula, this sample was adjusted for non-response, and an equal quota was allocated to
each of the neighbourhoods. For the experiment, an air quality monitoring station was
mounted in each of the neighbourhoods to gather data on SO2, O3, NO, CO, PM2.5, and
PM10 from 2017 to 2023. Upon a critical analysis of 455 papers in support of the conceptual
framework, key gaps were identified in the degree to which AI has been observed as a tool
to monitor and control air quality in dedicated smart city contexts. An even more acute gap
was observed in the number of studies that consider SDG goals to benchmark AI-air quality
research. Survey evidence supports the impact of meteorological factors and population
growth on air pollution, the impact of climate change and AI on air quality, and the impact
of AI and air quality on sustainable development. The moderating role of AI on air quality
and SDG outcomes was not statistically significant in the test for the role of smart city in the
model. AI experiment results also reveal that pollutant trends exist in seasonal patterns with
high correlations between them. Evidence also shows that SO2 is more dominant and that
the Al Jurf station has the highest levels of air pollution. As part of the AI experiment, the
artificial neural network (ANN) tool evolved as the optimal tool for air quality monitoring
and control based on the collated data. Following evidence from these three phases, data
validation is critically discussed in the areas of conceptual model specification, model
verification, and model validation. The present study makes an immense contribution to
knowledge as it is one of the first of its kind to consider AI for air quality monitoring and
control towards SDG air quality goals. A unique contribution to knowledge also exists as it
draws in the smart city context and employs an innovative methodology which includes an
SLR, survey research, and AI experimentation. It is recommended that such future research
consider traffic data, residents’ perceptions, and a case study of Al Jurf in the Emirate of
Ajman in an investigation into the use of ANN to improve smart city transportation through
air quality monitoring and control in contribution to sustainable development.