How the World Bank is using AI

Need help digging through environmental, social, and governance data for key insight, trends, or just information about where a company stands on its climate pledges? What about identifying the climate resilience of building projects? Or the climate impact of critical mineral supply chains? Turns out the World Bank has an artificial intelligence tool for that. While the world of artificial intelligence is changing fast, and the emergence of generative AI in the past couple of years has changed the landscape, the technology is not new for the World Bank, which has been using it for years, Harry Daniel Lersch, an AI strategist at the World Bank, told Devex. Generative AI — think ChatGPT, Google’s Gemini, or AI that can create new data content — emerged about two years ago and has been a “game changer,” he said, presenting “a very big opportunity with vast potential.” “The vision that we have for AI is that we use AI to accelerate, deepen, and sustain development impact, at the service of our vision and mission,” Lersch said. The global anti-poverty lender is ramping up how it uses AI both internally and externally. In terms of its own operations, AI is used by bank staff members to boost productivity and free up time previously spent on repetitive, routine tasks for more complex, value-creating activities, he said. That ties nicely into World Bank President Ajay Banga’s focus on improving efficiency and speeding up decision-making processes as part of broader efforts to reform the institution. To do so, the World Bank developed a proprietary internal AI tool called mAI embedded in the institution’s systems that relies on its data. It has similar capabilities to ChatGPT and allows staff members to easily search decades of knowledge, summarize content, search documents, and more. The World Bank also has an AI-driven tool that helps create outlines or early drafts for often-used documents, speeding processes, Lersch said. The bank’s use of AI is evolving as it learns and as the technology evolves, with the goal to always “use AI in a meaningful, effective, and responsible way,” he said. It has a framework that helps guide its work, including core principles around leveraging commercially available technology where it can, spurring innovation but also ensuring that proper safeguards around privacy and data protection are in place. “This is something that we would like to achieve, that we balance innovation and providing some level of safeguards and control,” Lersch said. “For us, it is also a learning journey, I guess it is for all organizations out there.” The bank, and Lersch’s team, is focused on training staff members on how to use AI, making sure that they apply the right data and technology and that they have appropriate safeguards in place to mitigate potential risks. AI is not completely fail-safe. Risks include the validity and accuracy of what AI generates — staff members need to verify data and use their expertise to judge its output — potential bias in the data, and users inputting sensitive information in external AI tools that are not authorized, Lersch said. The bank is also working on AI tools that can be used externally and is ramping up its work with governments on how they might use AI. The World Bank launched a GovTech Advisory and Guidance tool in 2023 that is meant to inform policy dialogue and design and implement future projects. Those could work on core government operations, service delivery, citizen engagement, and more. Several of the tools are focused on the bank’s private sector operations, including Malena, a public, free AI tool that analyzes environmental, social, and governance data that aims to speed analysis, increase productivity, and enhance investor confidence. It is also the only AI-powered solution tailored to emerging markets, according to the bank. It is meant to assist asset managers, banks, private equity firms, bank regulators, government ministries, and data providers to better make investment decisions and manage portfolios, improve oversight, and enhance data quality. There is also the Building Resilience Index, which gives the building or construction sector a tool to map hazards and a framework to assess the resilience of projects so they can be built to withstand climate change. The bank’s digital team is also focused on broader issues around global governance of technology, including AI, and the digital divide in low- and middle-income countries. That includes building the capacity of governments to understand, use, and regulate AI. The bank’s development impact group uses AI to improve impact measurement, but also to create new approaches to predict food insecurity, address gender biases, and advance conflict early warning systems. “As a bank, we have an obligation to also tap into this opportunity and explore how we can best use it at the service of our mission, and at the same time, we also want to make sure that we do this in an effective and responsible manner,” Lersch said.