Header image

4 Ways Government Can Unlock the Power of AI for Climate and Development

November 19, 2021 blog-post climate-change climate-action sustainability development government

At Thinking Machines, we are always looking for ways to apply AI and data science for development and social impact, including climate. The PwC reports that AI applications for the environment could drive global greenhouse gas emissions down by up to 4% by 2030. But too often, the public sector encounters challenges in digital innovation.

This same observation prompted the UN International Telecommunication Union (ITU), specializing in information and communications technologies, and other international groups bringing together climate action and AI like Climate Change AI (CCAI), the Centre for AI & Climate (CAIC), and the Global Partnership on Artificial Intelligence (GPAI) to release new reports this year compiling recommendations for governments to consider in using big data and AI for climate response and development.

While this recap of insights is far from exhaustive, here are four big challenges we hope the government and industry practitioners will take on:

Build affordable data infrastructures and interoperable data portals in sectors like energy and agriculture.

One of the common challenges we face when building AI applications for social impact is the lack of representatively sampled, disaggregated ground-truth training and validation data. And even if data is available, it is sometimes inaccessibly stored on on-premise legacy systems. These challenges delay and discourage progress in AI development.

The ITU report on Emerging technology trends: Artificial intelligence and big data for development 4.0 highlights the need for local data and adequate data infrastructures to support innovation. To complement this, the GPAI report on Climate Change and AI Recommendations for Government Action also recommends assembling data task forces in climate-critical sectors, establishing open data standards, and increasing data access and sharing through interoperable data portals. These task forces would identify gaps and opportunities in current data ecosystems and use their domain expertise to tailor-fit recommendations to specific sectors.

The GPAI’s AI-for-climate recommendations for government

What the Philippines is doing: There are multiple platforms for accessing historical and real-time data on climate-induced natural hazards. For example, there is GeoRiskPH for hazard and risk data by the Philippine Institute of Volcanology and Seismology (PHIVOLCS), GeoPortalPH for all base maps by the National Mapping and Resource Information Authority (NAMRIA), and Project NOAH for hazard assessments, housed in the University of the Philippines Resilience Institute (UPRI).

What can still be done: These portals, and many others, have overlapping data and functionalities, emphasizing the need for better coordination between data producers across public and private sectors.

Fund educational and professional programs that advance digital skills and bridge AI and climate-relevant sectors.

Developed countries should fund more AI-for-climate research and development in climate-vulnerable countries in the Global South, while national governments should also identify “grand challenges” where AI research and development can deliver the highest impact.

At the same time, digital literacy is becoming a core skill that governments need to strengthen. Citing the lack of knowledge within the public, private, and civic sectors on AI and climate, and the need for more cross-disciplinary and cross-sector experts, the GPAI report recommends rapid “upskilling” programs and funding higher education and professional opportunities that bridge AI and climate-relevant sectors.

What the Philippines is doing: The Philippine Department of Science and Technology (DOST) supports various AI projects, such as for the environment, transport, and supporting local government decision-making, and provides AI training for academic researchers.

What can still be done: Many of the projects funded are in partnership with universities, and there is room for stronger collaboration with private and civic sectors.

Develop policies that incentivize the experimentation, deployment, and scaling of solutions.

Many AI technologies and climate solutions get stuck in the early stages of development, because of a lack of financial incentives, data availability, slow adoption, and inherited legacy infrastructure. Governments can encourage investment to rapidly facilitate experimentation and bring local solutions from the pilot phase all the way to scale out.

What the Philippines is doing: As mentioned earlier, the DOST already supports AI R&D initiatives in the academic sector, from projects supporting agricultural monitoring to decision-support systems.

What can still be done: The DOST can work more with the private sector and civil society to create enabling environments for experimentation and incubation, and apply these innovations on the ground.

Implement standards, policies, and strategies that guide the inclusive and responsible use of AI.

AI can accelerate climate-positive solutions in every sector, but it can also have negative social and environmental impacts. AI can be used to accelerate natural resource extraction and consumption, amplify human biases, and exacerbate social inequality when trained on unrepresentative data. For example, a model trained on data from the US or Europe can’t be expected to perform the same in developing countries. And as earlier mentioned, AI literacy is also highly unequal. Models and outputs should be explainable to reduce barriers and increase trust and confidence, especially for non-technical policy- and decision-makers.

The ITU recommends governments develop AI and data strategies and roadmaps, by starting with identifying national challenges and priorities and conducting a SWOT analysis. Authors of the GPAI report also highlight that governments should engage a wide range of experts from different disciplines to implement ethical standards and best practices to ensure that AI positively impacts the climate-relevant contexts in which they are deployed.

What the Philippines is doing: Earlier this year, the Philippine Department of Trade and Industry (DTI) launched the national AI roadmap, supported by the DOST.

What can still be done: Given current gaps in AI policy and governance, the DOST reported plans to formulate AI laws and standards to increase security, transparency, accountability, and fairness.

The Philippine National AI Roadmap launched by DTI in 2021

For the full reports, read the ITU’s Emerging technology trends: Artificial intelligence and big data for development 4.0 here, and the GPAI’s Climate Change and AI: Recommendations for Government Action here.

We are eager to work with agencies in and working with the public sector to address climate and development challenges. If you think data or AI is a key component of your solution, reach out to us at [email protected]!


Creating a Single Customer View out of Big Messy Data

We worked with LBC to develop a Single Customer View (SCV) that automatically identifies unique customers from daily transactional records, resulting to 40x faster processing and 3.5x higher accuracy.

Let AI design your holiday cards!

Done making Christmas cards? Why not let an AI do it for you?

Thinking Machines’ Climate Action Commitment

Being data-driven means taking climate action. We’re trying to find our niche in the climate fight.