Leveraging Artificial Intelligence For Advancing Sustainability: Opportunities And Challenges

By Ernest Toh

Over the past few years, we have heard that sustainability is often closely linked with environmental issues. With the increasing threat of climate change, there is a rising demand for the effort required to ensure that today’s business development does not negatively impact our future generation.

The evolution of the environment and society, resulting from the Holocene (a relatively stable climatic period) to the Anthropocene (a proposed epoch marked by significant human impact on Earth’s systems), has impacted the Earth through human activity. The planetary boundaries framework sets up nine development processes for humanity, and we might face a large-scale change in Earth’s system if we go against it.

While governments are often seen as key contributors to achieving the United Nations (UN) Sustainable Development Goals (SDGs), what if their efforts alone are insufficient? Can meaningful progress be achieved if governments initiate programs but businesses and society fail to collaborate?

While efforts to achieve the SDGs may exceed global planetary boundaries, starting with small, achievable wins is essential. Business sustainability begins with evaluating a company’s value system and principles.

Artificial Intelligence And SDG

Companies can leverage emerging technologies to support sustainability, and artificial intelligence (AI) has become an especially notable tool. Below are a few examples of how AI is used today to achieve SDGs.

Climate Action

To combat climate change (SDG 13), DestinE—an initiative of the European Union—leverages AI with cloud computing, high-speed connectivity networks and big data to create a highly accurate digital model of Earth to monitor the effects of natural and human activity on our planet. This helps anticipate extreme events to ensure appropriate measures are taken to address climate-related challenges.

Justice For All

AI can help enhance the efficiency and effectiveness of the legal process (SDG 16). For example, the government uses AI to enhance national security (including trafficking) through facial recognition. Integrating the large language models (LLMs) within the context of national security helps to revolutionize information processing and decision making. The machine learning model can be used to accurately predict corruption patterns, for example, by using political and economic factors and leveraging a neural network approach to develop an early warning system.

Education

In education (SDG 4), we witness a surge in the use of AI-powered tools. According to Gartner, global AI software spending in education is expected to see a five-year CAGR of 18.5%, reaching an estimated $13.2 billion.

Educational institutions use AI to analyze student performance and identify personalized training needs. Teachers play an important role in the education sector. However, finding good quality teachers is not always easy, especially in developing countries. Adaptive machine learning models collect and analyze real-time student and teacher learning data, delivering immediate insights.

Healthcare

In healthcare (SDG 3), AI can manage vast patient data, enabling faster and more accurate diagnoses. The combination of big data and AI has enabled doctors to predict how likely a patient will develop an illness soon. This includes using AI to quickly identify acute stroke on CT scans to accelerate life-saving treatments.

Challenges Of AI

While AI has proven its worth in supporting the acceleration of the 2030 agenda for sustainability development, it poses challenges, too. Biases in data fed into the machine learning model affect decision-making and lead to undesirable consequences. For example, in healthcare, more data are collected in specific geographical locations (or age groups) than others, resulting in resources and effort in the wrong direction, affecting the overall sustainability goal.

The type of algorithm used affects the accuracy of the machine learning models, including the quality of the training data. The understanding of the business requirements and choice of the appropriate metric (e.g., accuracy versus recall versus precision) over another affects the cost and the time invested into the machine learning models. For example, recall is important when detecting illegal practices (SDG 16). We want every potential issue to be caught, and the goal here is to identify as many true positives as possible. The impact of missing even a small number of violations can have enormous consequences, including financial losses.

Ethical concerns are another rising factor when implementing an AI initiative. Personal data is often collected for model training without the user’s consent. In that case, individuals become less trusting and engage in AI systems, which impacts the adoption of AI to achieve a more sustainable outcome. Multiple countries have data confidentiality regulations, such as GDPR in the EU, PIPL in China and PDPA in Singapore to protect personal privacy. However, these also pose challenges to organizations in data collection and utilizing them.

Apart from aligning AI to meet the sustainability goal, there are other challenges to the AI implementation journey, which are more related to technicality and human factors. This includes integrating the model into the existing production system and ensuring it operates smoothly.

The dataset’s limitations and choosing the right model can be challenging. For example, decision trees can be more sensitive to noisy data, and if the data quality is poor, they may produce inaccurate predictions. Resistance to change from employees poses a risk to successfully rolling out AI solutions as well.

Conclusion

AI has enormous potential to support accelerating the 2030 agenda towards a more sustainable future. Through collective impact, the government, society and private sectors must unite to ensure data ethics, transparency, reliability and responsibility are considered in any AI initiative implementation.

Sustainability is a journey, not a destination—involving continuous learning, unlearning and relearning. This journey encourages us to see things from new perspectives and embrace a mindset shift. Gaining knowledge by uncovering the unknown is a crucial step forward.

To support accelerating the 2030 agenda for sustainable development, we must consistently challenge ourselves, leveraging and exploring current resources and technologies. Most importantly, this must be done without compromising the needs of future generations.