Whether future AI models are fair, trustworthy, and aligned with the public’s interests rests in part on our ability to collect accurate data about what we want the models to do. However, collecting high-quality data is difficult, and few AI/ML researchers are trained in data collection methods. The talk bridges artificial intelligence and survey methodology, demonstrating how techniques from survey research can improve training data quality and model performance. The talk concludes with practical recommendations for improving training data collection and ideas for joint research.