Harnessing AI to accelerate the Sustainable Development Goals

Featured guest post by Golestan (Sally) Radwan, Chief Digital Officer, United Nations Environment Programme

As we stand at the crossroads of technological innovation and environmental stewardship, AI emerges not just as a tool of transformation but as a catalyst for sustainable change. Its potential to advance the environmental dimensions of the Sustainable Development Goals is vast and varied. Through our work, the United Nations Environment Programme (UNEP) has identified five key domains where AI’s impact is particularly transformative.

Monitoring, Predicting Risk, and Creating Early Warning Systems

Firstly, through advanced algorithms and satellite imagery, AI systems can predict natural disasters, track biodiversity loss, and monitor climate change impacts in real-time and at a global scale. This predictive power enables nations to act swiftly and smartly, minimizing human and economic losses before they escalate.

One example is the International Methane Emissions Observatory hosted by UNEP, which employs AI to detect significant methane emission hotspots and facilitate timely mitigation actions. Additionally, the Flood and Drought Portal developed by UNEP-DHI Centre, aggregates and translates publicly available data from a range of sources, making it accessible to water authorities, helping them to mitigate water-related disasters.

Maximizing Resource Efficiency, Substitution, and Pollution Reduction

Secondly, AI-driven technologies optimize water, fertilizer, and pesticide usage in agriculture, and maximize efficiency in renewable energy production. Industries are applying these technologies to optimize resource use in a variety of ways including AI-powered smart devices to ensure efficiencies in heating, lighting, and energy use, significantly reducing the carbon footprint of homes and buildings. Simultaneously, these systems help minimize pollution and preserve our planet’s health by pinpointing sources and suggesting actionable solutions.

In Timor-Leste, for example, UNEP is testing how machine learning can be applied for the generation of crop advisories. The Specialized Expert System for Agro-Meteorological Early Warning (SESAME), a web portal, employs advanced algorithms to evaluate crop sensitivity to environmental conditions, assess the impact of weather on crop growth, and generate tailored advisories to promote resource efficiency.

Enabling a Circular Economy

Thirdly, AI helps enable a circular economy by supporting recycling processes, reducing waste, and encouraging the reuse of materials. A critical tool in this transformative approach is the AI-powered digital product passport, which provides detailed information about the materials and manufacturing process, enabling consumers to make informed choices and develop more sustainable consumption patterns.

Another example of this is a UNEP AI-based SDG Meter web platform which uses a neural, network-based technique for national language learning processing pre-training to help analyze and assess a text’s relationship to each of the 17 SDGs. This technique is based on the BERT (Bidirectional Encoder Representations from Transformers) model developed by Google researchers, with the use of the multilabel text classification feature. The results to date have been impressive, with an accuracy of 98% on 500, meaning that for 500 test texts the method correctly classifies 490 texts.

Empowering Green Consumer Behaviors and Lifestyles

Fourthly, AI can empower the adoption of green consumer behaviors and lifestyles. According to a study, 69% of respondents expressed concern about sustainability when it came to buying groceries, yet only 7% were actually buying sustainable products. Through apps and platforms such as UNEP’s Guidelines for Providing Product Sustainability Information in E-Commerce that analyze personal consumption patterns, or through recommendation engines, smart filters, and gamification, AI can improve transparency and suggest modifications to encourage individuals to make more sustainable purchases.

Bridging Environmental Science, Policy, Action, and Innovation

Lastly, AI identifies trends and solutions that inform policymaking and lead to better collaboration ensuring that scientific discoveries translate into actionable, impactful environmental policies and new technologies. This can also result in seamless data sharing and integration, allowing for faster analysis and implementation whilst simultaneously identifying sources of misinformation on the environment so they can be red-flagged and removed from prominent social media platforms. UNEP is already undertaking work in this area, testing a dedicated large language model which learns from authoritative and validated sources of environmental science as the basis for making policy recommendations.

Through increased collaboration, innovation, and implementation, together with the development of robust governance and ethical frameworks that include minimizing the environmental impact of AI itself, we must ensure that AI remains unbiased and benefits all. This is critical, because what is undeniable is AI’s potential to address the environmental challenges of our time and help steer our planet toward a more sustainable and prosperous future.