AI has been reinventing aspects of almost every industry, producing gains in productivity and contributing to economic growth. At the same time, it has been applied to solving some of the world’s most pressing problems.
A recent report from QuantumBlack, AI by McKinsey, revealed that the number of AI use cases aligning with the UN’s Sustainable Development Goals has grown to over 600—a threefold increase from 2018. Over 82 percent of these AI solutions have been deployed, demonstrating the rapid pace of innovation and adoption in this field; however, most are still in early pilot phases.
McKinsey has been helping develop AI solutions in two areas: job placement and disaster relief efforts.
Accelerating job matching with AI
In 2015, McKinsey launched Generation, an independent nonprofit that trains and places adults in careers they would typically not be able to access. Beginning as a small organization, Generation now operates in 17 countries with more than 120,000 graduates since its launch.
Generation staff members source employment opportunities by scouring job search sites for over 30,000 graduates each year. This labor-intensive task averages five to six hours weekly for around 100 people—time that could be spent coaching and supporting learners.
“Generation wanted to automate this process,” explains Oskar de Smet, an associate in McKinsey’s New York office. “In around three months, we built a working prototype of an AI-enabled tool that looks at learners’ profiles and vacancies and then matches them.”
The first step was creating a taxonomy of skills: “Basically a long list of skills you’d see in a vacancy ranging from specialized technical skills to soft skills like communication,” says Oskar. “Then we go through learners’ CVs—or they can self-report online—to collect their top skills.”
“When a cohort is ready to graduate from Generation’s program, the tool can help match vacancies from different platforms with skills in our taxonomy,” explains Oskar. Learners then receive a list of potential openings, rated for relevance—helping graduates find better jobs faster and improving career outcomes.
Ensuring responsible AI practices was also critical, and the team mitigated risks related to impaired fairness, data privacy, and IP infringement, and verified compliance with global data regulations like the GDPR.
The tool has proven successful in pilots in the UK and Mexico, and Generation plans to roll it out to all 17 countries starting in January 2025.
“Our partnership with Generation has grown from building foundational capabilities and tools to the latest innovations in digital and AI,” says Prasoon Sharma, a partner at McKinsey.
Focusing disaster relief efforts with data
To use AI capabilities to improve disaster relief, McKinsey joined forces with Google, the Jain Family Institute, and the Patrick J. McGovern Foundation to found DISHA (Data Insights for Social and Humanitarian Action), an initiative led by the UN Secretary General’s Office of Innovation.
This effort is part of Noble Intelligence, McKinsey’s AI for Social Good initiative, which applies our QuantumBlack data and AI capabilities to pro bono social impact projects.
QuantumBlack, AI by McKinsey
Via QuantumBlack, McKinsey provides strategic and technical expertise to DISHA. Our data scientists and engineers support the UN in building and refining technical solutions using responsible AI practices. We also aid in roadmap development and offer training on how to use the solutions, interpret data, and integrate the tools into business processes.
DISHA has developed two solutions to make responses to natural disasters faster and more targeted.
The first, which addresses damage assessment, uses machine-learning-based tools and satellite image analysis to map out damage to buildings and infrastructure. This tool will allow the UN and its partners to analyze areas seven times larger while reducing the time required for damage assessment findings sixfold—bringing it to under a day. This speed and scale will enable humanitarian organizations to more quickly mobilize resources to communities.
The second solution is socio-economic mapping, which helps humanitarian organizations estimate poverty levels and identify the people most in need. Traditionally, humanitarian responses have relied on outdated data from censuses and house-to-house surveys to figure out where to send aid after a natural disaster. DISHA’s solution uses real-time data from mobile phone operators to track where people are moving on a local level, so organizations know immediately where support is most needed. This technology is already active in the Philippines, where thousands of villages are regularly affected by typhoons.
“Our approach is unique in that it unites diverse partners to share data, build trust, and create digital public goods,” explains Ankit Bisht, a partner at McKinsey. “DISHA’s impact extends beyond just a solution; it empowers future generations. We are developing new frameworks for collaboration between public, private, and non-governmental organizations and creating models to share data for social good that follow responsible AI practices.”