1144: Deep Learning and Collaborative Evaluation: Tools, Ethics, and Opportunities in the Age of AI

As digital transformation reshapes programs, policies, and services, evaluators increasingly encounter complex, high-volume data and novel sources of insight. Deep learning—a branch of artificial intelligence (AI)—offers tools for pattern recognition, predictive modeling, and automated data processing that can enhance evaluation practice. A collaborative evaluation lens offers a crucial bridge, helping evaluators interpret these tools in context, promote shared understanding, and prioritize community values.
This preconference training introduces deep learning through a collaborative evaluation framework. Designed for professionals with limited or no technical background, the session explains concepts such as neural networks, model training, supervised versus unsupervised learning, and interpretability. Participants will examine case examples from education, health, and social impact evaluations—such as text analysis, image classification, and time-series prediction. The session includes interactive demonstrations and group activities (no coding required), with equity, transparency, and inclusion integrated throughout. Participants will gain the tools to assess deep learning methods, co-interpret results with stakeholders, and lead ethical, collaborative conversations about data.

Michael Harnar; Michelle Rincones-Rodriguez