Advances in data and measurement are transforming how researchers and policymakers understand jobs, livelihoods, and economic resilience in low- and middle-income countries, but improving survey quality remains complex. The way questions are designed, responses are recorded, and new technologies are integrated can significantly influence the accuracy of results. These issues were central to the inaugural conference “Better Data for Better Jobs and Lives: Innovations in Survey Measurement in the Age of AI,” held at the World Bank in December 2025, which brought together researchers working on strengthening survey methods in an era of rapid technological change. Building on this momentum, a follow-up conference is scheduled for December 2026 in Washington, D.C., focusing on continued innovation in survey measurement.
A key message from the discussions is that household surveys remain the foundation of development data, even as AI and new data sources expand analytical possibilities. They continue to provide essential ground truth for understanding poverty, employment, food security, and resilience, while also helping validate emerging tools such as machine learning models and geospatial analytics. At the same time, the conference highlighted that measurement choices—such as question design, survey mode, and sampling approach—can significantly alter results, sometimes producing variations comparable to the effects being studied. This shows that even small methodological differences can meaningfully change policy-relevant findings.
The discussions also emphasized that measuring jobs and livelihoods in low- and middle-income countries requires approaches tailored to informal and complex labor markets. Unlike high-income economies, many workers combine multiple income sources, shift between self-employment and wage work, or operate in informal settings that standard surveys often miss. Researchers presented new methods to better capture these realities, including improved sampling strategies, alternative data collection techniques, and more detailed indicators of employment quality beyond wages and job status.
Another major takeaway was the growing role of high-frequency data and new technologies in expanding what surveys can measure. More frequent data collection is revealing short-term changes in income, employment, and time use that traditional surveys often overlook. In addition, tools such as satellite imagery, sensors, and machine learning are being used to complement surveys and fill data gaps, particularly in areas where traditional data collection is limited or infrequent.
The conference also highlighted how artificial intelligence is already improving survey processing and analysis. AI tools are being used to classify jobs, code open-ended responses, and analyze qualitative data more efficiently, reducing costs and enabling faster insights. These technologies are also helping researchers harmonize datasets and extract value from unstructured information that was previously underused. However, discussions stressed the importance of ensuring methodological rigor, transparency, and appropriate safeguards when applying AI in survey workflows.
More experimental AI applications, such as conversational interviewers and voice-based analysis tools, were also discussed, but most remain at early testing stages. While they show promise for improving data collection and analysis, concerns remain about accuracy, bias, and reliability. A central message was that AI tools must be held to the same standards as traditional survey methods before being widely adopted.
Finally, the conference underscored that the biggest challenge is not innovation itself, but scaling and integrating new methods into real-world survey systems. Strengthening data systems will require not only technological advances but also institutional capacity and skills development. Despite rapid progress in AI and alternative data sources, high-quality household surveys remain irreplaceable for capturing detailed and context-rich information. The overall goal remains improving measurement systems to generate better, more reliable data for policy and development outcomes.
