Artificial Intelligence & IoT Driven Technologies for Environmental Pollution Monitoring and Management

Introduction:Hazardous substances in the environment are those that pose a threat to human health, plant, and animal life, or the environment. These substances include heavy metals, pesticides, herbicides, and persistent organic pollutants (POPs) that have been introduced into the environment through various means. Sources of hazardous substances in soil include industrial activities, improper disposal of hazardous waste, agricultural practices, and natural processes such as erosion and weathering. These substances can persist in the environment for long periods and can harm soil quality, plant growth, and human health. The effects of hazardous substances in the environment can vary depending on the type and concentration of the substance, as well as the duration of exposure. Some hazardous substances can cause acute health effects, such as respiratory problems, skin irritation, poisoning, nausea, and vomiting, while others can lead to chronic health problems, including cancer, reproductive disorders, and developmental abnormalities.Effective management of hazardous substances in the environment requires monitoring, remediation, and prevention strategies. Monitoring involves regular testing of soil for the presence of hazardous substances, which allows for early detection and appropriate management strategies to be implemented. Remediation involves the removal or treatment of contaminated sites. Prevention strategies include reducing the use of hazardous materials (i.e., source control) and implementing best practices for waste disposal and land use. Effective management is essential for ensuring the long-term health and sustainability of soil and its ecosystems. Real-time monitoring of hazardous materials in soil and plants is an important task that can help to ensure the safety of food crops, protect the environment, and prevent human exposure to harmful substances. The use of artificial intelligence (AI) powered sensors and devices can greatly enhance the accuracy and efficiency of this monitoring process. AI-powered sensors and devices can be used to detect and quantify the presence of various hazardous materials in soil and plants. These sensors and devices can be designed to measure parameters such as pH, temperature, moisture, conductivity, and various chemical properties of the soil and plant tissue. Machine learning algorithms can be used to analyze the data collected by these sensors and devices, enabling the identification of specific hazardous materials in real time. These algorithms can also be used to predict the potential impact of these materials on human health and the environment. For example, E-nose (olfactory) algorithms are used to analyze data generated by sensors and identify the presence of hazardous chemicals based on their unique chemical signature. These algorithms can use a variety of techniques, such as pattern recognition, artificial neural networks, and fuzzy logic. One of the key advantages of E-nose technologies is their ability to detect hazardous chemicals in real time, allowing for immediate response to potential threats. E-nose technologies can be used for a variety of applications, such as monitoring air quality in urban areas, detecting leaks from industrial processes, and detecting explosives and other hazardous materials.The use of AI-powered sensors and devices for real time monitoring of hazardous materials in soil and plants has several benefits. Firstly, it allows for more accurate and reliable detection of these materials compared to traditional laboratory-based methods. Secondly, it provides real-time data, allowing for quick responses to any potential contamination events. Finally, it reduces the need for manual data collection and analysis, reducing the workload and increasing the efficiency of the monitoring process. The use of AI-powered sensors and devices for real time monitoring of hazardous materials in soil and plants is a promising approach that can help to ensure the safety of food crops, protect the environment, and prevent human exposure to harmful substances. Various approaches can be used for AI based toxicity prediction, including machine learning methods, deep learning methods, and hybrid approaches that combine both methods. Integrating various sources of data, such as chemical structures, toxicological and physiological data, and environmental factors, to improve the accuracy and reliability of toxicity predictions, is important.One of the recent advances is the combination of AI and Internet of Things (IoT) technologies, for particulate matter (PM) monitoring, which uses low-cost sensors that can be easily deployed in various environments. These sensors can collect data on PM levels and send it to a centralised platform for analysis. AI algorithms can then process these data to provide real-time information on PM levels and predict future trends. Heavy metals can also be monitored using AI. Numerous studies have been conducted over the last 10 years to forecast the effectiveness of heavy metal removal from soil using machine learning. AI models for the optimisation and prediction of heavy metal removal include black box, fuzzy logic, kernel, evolutionary, and hybrid models. With the rapid changes that the environment is experiencing, data sharing and reuse with the help on AI algorithms and instruments, plays an important role in supporting researchers to safeguard the continuously threatened environment and ensure the implementation of sustainable environmental management practices. Scientists can make use of online data sharing tools and platforms that comprise vast and intricate Earth and environmental science data like climatic and atmospheric data, pedology, hydrology, ecology, and biodiversity data for testing, analysing, interpreting theories, prediction models and experimental data that lead to better understanding environmental issues..of heavy metal removal include black box, fuzzy logic, kernel, evolutionary, and hybrid models. With the rapid changes that the environment is experiencing, data sharing and reuse with the help on AI algorithms and instruments, plays an important role in supporting researchers to safeguard the continuously threatened environment and ensure the implementation of sustainable environmental management practices. Scientists can make use of online data sharing tools and platforms that comprise vast and intricate Earth and environmental science data like climatic and atmospheric data, pedology, hydrology, ecology, and biodiversity data for testing, analysing, interpreting theories, prediction models and experimental data that lead to better understanding environmental issues.In this article, we will examine the application of artificial intelligence (AI) in monitoring hazardous materials across different environments, namely, soil, air, and water. We will explore the latest breakthroughs and progress made in this f ield, which integrates machine learning algorithms with sensor technologies. We will also consider the benefits and drawbacks associated with AI-powered monitoring systems. This new field has revolutionised soil, air, and water monitoring, enhancing accuracy, efficiency, and timeliness in detecting and analysing hazardous substances..In recent years, there has been an increase in interest in using AI to anticipate and predict environmental pollution. We can split AI solutions into three steps or phases that include inputs (data), models (AI algorithms), and outputs (monitoring or decision support) (Figure 2). Data forms the basis of any AI solution. These models work best when they have a high number of data points, especially the ones that are coming from environments that are continually being monitored and expanding solutions for probable actions. Data sources today can be a variety of sensors, ranging from imaging to non-imaging types or remote to in contact sensors that provide large volumes of data. Then there are the historic or legacy data. AI algorithms can analyse massive amounts of sensor readings, historical data, and other important information from monitoring systems. AI can detect hazardous material levels and contamination events by identifying patterns, trends, and anomalies in data using machine learning and data mining. Image analysis can detect hazardous material spills via satellite photos or drone images. Computer vision algorithms can recognise chemicals, vegetation changes, and pollution sources.AI algorithms can be trained on past data to predict hazardous material releases and environmental pollution. AI solutions for environmental monitoring, thus, would include early warning systems for hazardous material release, autonomous pollution monitoring systems as well as decision support systems. These models can help authorities and organisations prepare for and respond to emergencies. These models have proved useful for multiple environmental conditions, be it soil, air, or water. AI modellers, however, should offer sufficient details to explain and support the selection of model parameters, as well as their creation and assessment..Many short- and long-term forecasting applications use ANNs (artificial neural networks). One has created an IoT-enabled environmental toxicology model to detect air pollution. The model uses artificial intelligence to report the status of the quality of air in real-time utilising a cloud server and broadcasts alarms when hazardous pollutants are present. The AAA (artificial algae algorithm)-based ENN (Elman neural network) model classifies and predicts air quality in future timestamps (a timestamp is the current time of an event that a computer records). WiFi gateways send the data collected by sensors to a cloud server. AAA optimises the ENN model parameters during data processing. For monitoring and enforcement, mobile electronic-nose (E-nose) devices and algorithms have been developed to quickly detect pollution from point sources. Due to their sensitivity, E-nose devices can detect carbon, GHG emissions, and particle pollutants emitted into air or water.Traditional soil mapping entails physically collecting soil samples and transferring them to a laboratory for additional analysis. The advent of technologies permitting high-resolution, quick, and inexpensive mapping of soil pollutants has favourable aspects over traditional techniques. Jia et al. (2021) created an unique modelling method that forecasts soil arsenic levels using high resolution aerial imagery (HRAI) photos.The method makes use of cameras that are installed on aircraft to take high-resolution (0.1–0.5 m) pictures of broad areas. The f irst layer of a model for displaying soil arsenic levels shows a thorough report on soil contamination and HRAI. The image is broken down into pixels in the second layer, and sample points are represented by pixel features. To forecast the risk levels of arsenic, four alternative machine-learning algorithms were constructed. The Extreme Random Forest (ERF) algorithm had the best prediction and accuracy. Remote sensing and aerial imageries provide continuous spatial data which, coupled with machine learning models, are providing highly accurate maps of hazardous substances in the environment that were not possible with standard geostatistical techniques..Aerial Imaging and Unmanned Aerial Vehicles (UAVs)Since the early 1990s, aerial imagery has been employed to monitor hazardous waste. Aerial photographs have proven useful in various applications for detecting and analyzing the presence of hazardous waste, waste-disposal sites, and landfills. Historical aerial photographs provide valuable documentation for compiling a record of changes occurring at hazardous sites and are considered reliable for monitoring changes over time. The U.S. Environmental Protection Agency (USEPA) has made use of an aerial imagery library that dates back to the 1930s to retrace the history of waste management and disposal at hazardous waste sites. Environmental clean-up programmes have generated more than 4,000 historical aerial photographic reports on hazardous waste activity and employed them in clean-up.Aerial photography allows for the interpretation of various objects at hazardous waste sites, including evidence of discarded materials, barrels and drums, open dumps, spills, and disturbance. These features enable monitoring and analysis of potential impacts related to hazardous waste. Aerial imagery facilitates the identification of vegetation patterns, investigation of local groundwater movement to assess potential pollutant migration, determination of drainage routes, examination of hydrological conditions, and evaluation of subsequent land use on closed landfills.Hyperspectral imagery-based models have proven effective in predicting heavy-metal distribution in soils. In a study conducted by Tan et al, a Competitive Adaptive Reweighted Sampling (CARS) method was proposed for this purpose. The researchers compared the accuracy of various models and discovered that CARS in combination with a stacking method exhibited the highest accuracy and stability. This method utilised the spectrum in the range of 2–2.3 μm, which is a common characteristic band for heavy metals. Additionally, it effectively addressed challenges, such as overfitting due to imbalanced data and limited training sample sets. Importantly, even in the presence of spatial heterogeneity, the distribution of heavy-metal concentrations derived from the CARS-stacking method showed consistency in the verification analysis.Increased population and industrialisation have led to a rise in hazardous waste spillage or leakage incidents. Detecting and monitoring toxic, flammable, and inert gas leaks are a global priority and have been the subject of extensive research. The emergence of Unmanned Aerial Vehicles (UAVs) or drones has significantly enhanced the collection of aerial imagery. This technological advancement has provided a boost in obtaining high-resolution and up-to-date aerial data for various applications, including the monitoring and analysis of heavy-metal distribution in soils. Forward reconnaissance drones, equipped with sensors, are valuable tools for enhancing situational awareness in environments that are challenging for humans. Remote gas detection systems, utilising mobile robotic platforms such as drones, have emerged as a promising approach in this field, especially in environments too hazardous for human exploration.By augmenting drones with sensors, hazardous materials or spills can be identified, and waste can be monitored to enable effective planning and management, thereby minimising the exposure of f irst-response teams. Also, integrating sensor-based particulate detection with autonomous drone flight control enables dynamic identification and real-time tracking of aerial plume boundaries. The findings of Seiber et al. (2018) demonstrate that UAVs can precisely recognize and track contaminant plumes over time, thus providing visual indicators and aid data collection that can be used to validate and advance the plume movement models.In synopsis, the utilisation of sensor-equipped drones for forward reconnaissance purposes can significantly contribute to identifying and monitoring hazardous materials, spills, and gas leaks in environments where human access is unsafe. The integration of sensors and autonomous flight control enhances the capabilities of drones in real-time plume tracking and provides valuable data for improved modelling and management of hazardous incidents.In incidents that involve hazardous materials, drone flights play a crucial role in promptly and accurately identifying the direction of spillage or gaseous material dispersion. Numerous studies have highlighted the ability of UAVs to be rapidly deployed to disaster sites and gather essential data about the extent and impact of spills. This information is vital for effective planning, management, and response efforts.Integrating UAVs into the toolkits of industry and government spill response teams can greatly enhance response capabilities, mitigation strategies, and overall accountability. In summary, UAVs are indispensable in assessing and managing incidents involving hazardous materials. Their ability to be quickly deployed, gather data, and monitor critical areas provides valuable insights for response teams, enabling them to make informed decisions, mitigate risks, and safeguard both human lives and the environment.Satellite remote sensingRemote sensing is a scientific and technological approach that enables the identification and assessment of various characteristics and properties of Earth’s targets from a distance. On both a global and local level, it has offered systematic and repeated observations of numerous features of the Earth’s surface, including the atmosphere, water, land, living things, vegetation, pollution, and climate. The utilisation of remote sensing has played a crucial role in detecting and quantifying pollution rates, mapping, monitoring, and mitigating pollution.One of many applications of remote sensing is monitoring and managing hazardous waste and sites. Multispectral sensors (MSS) mounted on remote sensing platforms are capable of digitally collecting energy levels of reflectance in specific bands of the electromagnetic spectrum. These systems offer advantages such as statistical analysis of data and the ability to extend observations beyond the capabilities of aerial photography. Land use, regional risk assessment, and hazardous waste site spectral characteristics and pollution profiles have been monitored using multispectral imaging systems mounted on satellites and various aircraft-based systems. For instance, Bølviken et al. demonstrated that MSS data could be employed to identify heavy metal contamination based on fundamental spectral characteristics.Furthermore, multispectral imagery has been used by several researchers to identify and locate previously unknown as well as illegal hazardous waste sites. Errico et al. proposed a methodology that combines synthetic aperture radar (SAR), multispectral data, and GIS-based processing for detecting environmental hazards.This system yielded satisfactory results and contributed to countering and controlling environmental crimes. Additionally, hyperspectral imagery can be employed to identify and map the spatial distribution of various heavy metals. For instance, Kemper and Sommer used a HyMap sensor and utilised airborne hyperspectral images to map lead and arsenic contamination in the Guadiamar flood plain, Andalusia. Similarly, Wu et al. (2011) reported satisfactory outcomes in heavy metal mapping in Nanjing City, China, using simulated HyMap data. They reported that direct predictions based on hyperspectral images often require signals from bare soils, which can be achieved during winter or early spring or when there is low vegetation coverage due to crop rotation in agricultural areas. Therefore, remote sensing has been instrumental in monitoring and managing hazardous waste and sites..Nowadays awareness of the safety of employees working in the production and manufacturing operations in different industrial environments that involve hazardous substances and materials has become of utmost interest. Thus, monitoring the concentration and leakage of pollutants and development of safe and efficient methods to reduce the exposure of humans and the detrimental risks for the environment, with fewer accidents provoked by hazardous substances, are becoming a priority in the implementation of safety plan management in many sectors of the economy.In pollutant discharge events, immediate, precise, and intelligent intervention is needed to alarm, prevent, and control hazardous leakage. Since humans cannot identify in time possible threats, because the majority of gases are odourless, colourless, and tasteless, implementation in environmental monitoring and detection of artificial intelligence based devices is getting more popular. The popularity of AI and unmanned machines arises from the effective, real-time, automated solutions that these technologies provide when placed in hazardous environments. They can detect and isolate possible threats before they cause harm, with less or no human involvement.Recent advances have been developed for the combination of autonomous robots with sensors that are flexible and easily and remotely deployed in unsafe and toxic environments. They can improve the safety of workers and neighborhood areas without the risk of involving human life. These devices are embedded with systems capable of exploring, monitoring, detecting, and alarming, in case hazardous events occur. For example, Joshna et al. developed an independent robot using Arduino UNO (R3), equipped with sensors that can detect and identify toxic gases. It can also perform degasification making use of oxidizing agents that are sprayed on the gases, reducing their deleterious effects.Fan et al. (Fan et al., 2019) developed a mobile robotic system equipped with an E-nose that is highly accurate in detecting and identifying different toxic gases like CO and NO2 present in the environment. It can be used in fire and emergency response departments. Gallego et al. (2015) developed an unmanned aerial vehicle (UAV) equipped with metal oxide semiconductor (MOX) gas sensors, capable of detection of toxic gas leaks, to monitor gas pipes in outdoor areas. Novelty traits of the system include improved velocity and reduced energy consumption and investment costs.Over the years, pollution of water resources has increased, making uncontaminated water bodies even more scarce all over the globe. The pollution has destroyed aquatic ecosystems and resulted in health and safety issues for humanity. Protecting water resources and preventing the release of different pollutants represent the first steps in ensuring safety for the environment and people. Various approaches for monitoring the physico-chemical parameters of water resources have led to the implementation of novel smart technologies. Water-quality smart monitoring systems using the Internet of Things (IoT) have gained popularity by engaging WSNs that offer continuous, remote and real-time monitoring and measurement of different water parameters. They can identify pollution sources and give the opportunity of early warnings when possible threats or hazards are detected, and they can provide proper safety measures.A low cost system making use of the IoT for the detection of water parameters in real time, like temperature, turbidity, conductivity, pH, and dissolved oxygen, which uses Raspberry PI B+ as a core controller, has been designed. Subsequently, Pasika and Gandla proposed a low cost, efficient system based on the IoT using ThinkSpeak and various sensors to monitor drinking water quality in real time. Similarly, Kumer et al. used ThinkSpeak and an Arduino Mega 2560 microcontroller in a device to analyze real-time water parameters to detect levels of contamination when it happens and to warn inhabitants of hazardous health risks. It is very accurate and inexpensive.Khan et al. (2020) developed a system that can monitor water quality for an industrial effluent treatment plant. It is composed of wireless sensor networks, a GSM module for notifications in case of emergency, an Arduino Uno R3 microcontroller, and an IoT-based cloud server. This system proved to be efficient and cost-effective. It gives continuous, real-time water quality monitoring, which allows the local authorities the benefit of supervising and checking to see if the released water from a specific industry is polluted or not. To tackle the problem of unreachable water bodies that are in secluded areas, Vasudevan and Baskaran (2021) proposed an innovative unmanned surface vehicle (USV) based on the IoT and sensors that provided control of these waters and diminished water pollution. They confirmed its efficiency, low investment, reduced energy consumption, and real-time data acquisition for surface water quality monitoring. Similarly, Ryu (2022) proposed an unmanned aircraft system (UAS) for water sampling and monitoring, which can be used in potentially hazardous environments, minimising the involvement of humans. Its UASWQP platform can take water samples from various points and is effective in presenting real-time data regarding measurements of water parameters..Khan et al. (2020) developed a system that can monitor water quality for an industrial effluent treatment plant. It is composed of wireless sensor networks, a GSM module for notifications in case of emergency, an Arduino Uno R3 microcontroller, and an IoT based cloud server. This system proved to be efficient and cost effective. It gives continuous, real-time water quality monitoring, which allows the local authorities the benefit of supervising and checking to see if the released water from a specific industry is polluted or not.To tackle the problem of unreachable water bodies that are in secluded areas, Vasudevan and Baskaran (2021) proposed an innovative unmanned surface vehicle (USV) based on the IoT and sensors that provided control of these waters and diminished water pollution. They confirmed its efficiency, low investment, reduced energy consumption, and real-time data acquisition for surface water quality monitoring. Similarly, Ryu (2022) proposed an unmanned aircraft system (UAS) for water sampling and monitoring, which can be used in potentially hazardous environments, minimising the involvement of humans. Its UASWQP platform can take water samples from various points and is effective in presenting real-time data regarding measurements of water parameters..BeeTox AI and other modelsIn a recent study, a novel web application called BeeToxAI was introduced. It utilises AI to assess the acute toxicity of chemicals to Apis mellifera, commonly known as the honey bee. BeeToxAI offers users a comprehensive set of features, including the classification as toxic or non-toxic of acute contact toxicity and acute oral toxicity endpoints. In addition, confidence scores for the prediction are given along with a visual representation of the results through maps. These color-coded maps illustrate the relative contribution of chemical fragments to toxicity. Since pesticides contain toxic substances that are injurious to health, AI algorithms are needed for controlling and tracking them, to reduce their toxicity. Because nitrogen based chemical fertilizers increase the amount of nitrate in groundwater, AI might be used to control excessive nitrate in soil.The analysis of water quality as a source of irrigation can be done using AI-ANN models (Ostad-Ali-Askari et al., 2017). Over the past decade, a multitude of studies have been conducted to forecast the efficacy of machine learning in removing heavy metals from soil (Table 2). These investigations have focused on predicting the effectiveness of various techniques for soil remediation and heavy metal removal utilizing machine learning methodologies. Models used for optimising and predicting heavy metal removal have encompassed black box, fuzzy logic, kernel, evolutionary, and hybrid models, and each offers distinct advantages.In a study conducted by Talebkeikhah et al. (2022), which focussed on the adsorption of Pb (II) on biochar, eleven models were employed, which included group-data handling methods such as support vector machines (SVM), radiofrequency (RF), adaptive neuro-fuzzy interference systems (ANFIS), multilayer perception (MLP), and Decision Tree. Additionally, to evaluate the significance of coupling in creating predictive models for estimating adsorption efficiency, the researchers developed four coupled predictive models using the grasshopper optimisation algorithm (GOA) and the bat algorithm, and they combined ANFIS and MLP..Chen et al. (2020) developed a CNN architecture specifically designed for deep calibration using near-infrared (NIR) spectroscopy data. Their study focused on assessing the level of water pollution originating from domestic and industrial sources, with the aim of enabling suitable agricultural irrigation practices. The researchers successfully established intelligent spectroscopic models using the CNN architecture, which could be instrumental in addressing water recycling and conservation issues in agricultural cultivation..Industry 4.0 is revolutionising the way companies manufacture, improve, and distribute their products, and AI is widely recognised as a crucial technology for it (Mao et al., 2019). Its incorporation in these domains has significantly transformed and enhanced various aspects of industrial processes, enabling advanced automation, data-driven decision-making, predictive maintenance, and optimisation of manufacturing operations (Figure 3). AI’s capabilities have played a pivotal role in driving efficiency, productivity, and innovation in the manufacturing sector, making it a key enabler of the Industry 4.0 paradigm. Industry 4.0 has brought about many cutting-edge technologies, and this new phase of AI development is known as AI 2.0 (Pan, 2016). The use of AI in smart factories and its application in modern industrial sectors have been acknowledged as transformative.It enables efficient decision-making processes based on real-time and historical data, minimising the need for human intervention. AI, machine learning, and autonomous technologies have significantly impacted the mining industry by detecting hazards and reducing risks. The introduction of autonomous trucks and the utilisation of AI and machine learning have brought cost savings, increased productivity, improved worker safety, and continuous production.AI systems analyzing geological, topographical, mineralogical, and mapping data can identify anomalies, locate potential areas of interest, and facilitate autonomous drilling operations Advancements in artificial intelligence (AI) and machine learning (ML) have significantly improved efforts in mitigating pollution. State-of-the-art technologies are being used to improve air quality modeling, integrating a variety of data sources such as satellite imagery and meteorological data for more accurate predictions.The amalgamation of AI with satellite technology allows for comprehensive monitoring of environmental shifts, assisting in the identification of pollution sources. The introduction of edge computing enables real-time analysis of environmental data at its origin, facilitating swift responses to pollution incidents. Hybrid models, which merge physics-based simulations with ML, enhance the precision of identifying pollution sources. ML-driven predictive analytics forecast pollution levels, enabling proactive measures to prevent environmental deterioration..While AI brings forth numerous benefits, it is not extempted from challenges and limitations that require efficient and notable solutions. These issues encompass the effective execution of tasks and management and mitigation of risks, as well as the assessment of hazardous materials. The requirement for significant processing power raises concerns about environmental impact and energy consumption. Developing countries may face infrastructure limitations that hinder the efficient deployment of AI. Data acquisition costs in data-driven agriculture remain high, limiting the widespread impact of AI on agricultural productivity. Furthermore, issues of data ownership, privacy, and cybersecurity arise, necessitating standardised protocols and integration with technologies like block chain to ensure data integrity and security. Block chain is a system in which a record of transactions is maintained across computers that are linked.In the manufacturing field, there is a need for further research on the application of AI in various sub-components, such as design, machinery used, assembly, and production. Ethical considerations also arise, including concerns about genetic engineering, moral decision-making by AI systems, and the impact on jobs, economy, and society, including privacy under surveillance.The AEC industry faces challenges related to funding, data, ethics, privacy, trust, scalability, and expertise. Resistance to technology adoption stems from concerns about job displacement, uncertainty, wealth distribution, and the complex relationship between humans and technology. Additionally, there are challenges in understanding how AI can effectively support individual and group decision-making. Overall, the application of AI in various industries brings about significant opportunities and challenges, necessitating further research, infrastructure development, and ethical considerations..The development of artificial intelligence-driven sensors for environmental monitoring of hazardous substances has the potential to revolutionise how we detect and respond to environmental threats. With their ability to process large amounts of data in real-time and identify patterns and anomalies that may indicate the presence of hazardous substances, these sensors can greatly improve our ability to protect public health and the environment. However, there are still many challenges that must be addressed, such as ensuring the accuracy and reliability of these sensors and determining the best ways to integrate them into existing monitoring systems. Despite these challenges, the potential benefits of AI-driven sensors make them a promising avenue for future research and development in the field of environmental monitoring. In the near future, AI-powered hazardous substance monitoring has the potential for enhanced automation, the creation of more sophisticated models, and the fusion of AI with novel technologies like the IoTAI could deliver promises about prediction, optimisation, and decision-making soon to help the traditional construction industry keep up with the rapid speed of automation and digitalisation in a variety of risky and dangerous circumstances. It can offer a wide range of risk identification factors as well as assess and deliver for prioritisation, because it can monitor, recognise, analyse, and anticipate possible risks in terms of safety, quality, efficiency, and cost across teams and work areas even in the presence of significant uncertainty. AI-powered monitoring systems can provide real-time, on-ground detection of potentially dangerous substances and the location of the incident sites. They may prevent any serious accidents from happening that would be harmful to human health. These systems are applicable in the fields of agriculture, health, mining, manufacturing, and other industries.real-time analysis. This can be achieved through the integration of sensors, data processing capabilities, and machine learning algorithms. By leveraging these technologies, AI systems can swiftly and accurately analyse data streams to identify potential hazards. This application is precious in industrial settings or areas susceptible to chemical spills or leaks. Furthermore, AI can contribute to predictive modelling and risk assessment of hazardous substances. By examining historical data and considering various environmental factors, AI algorithms can forecast the probability and severity of incidents involving hazardous substances.This allows for proactive mitigation measures to be implemented. AI can also integrate data from multiple sources, such as sensors, satellite imagery, and public health records. By combining diverse datasets, AI algorithms provide a holistic understanding of the distribution, movement, and impacts of hazardous substances. In addition, AI can offer decision support tools for human operators engaged in hazardous substance monitoring. Through user friendly interfaces and natural language processing, AI systems can aid in data interpretation, anomaly detection, and timely decision-making. These features assist operators in effectively managing hazardous substance situations