Northwestern Launches Agentic AI Challenge for Investigative Journalism

Northwestern University's Medill School of Journalism and McCormick School of Engineering have announced the Agentic AI Investigative Journalism Challenge, officially opening on May 15, 2026. The initiative invites cross-disciplinary teams—including journalists, developers, students, researchers, and civic technologists—to design and build reusable AI-assisted investigative workflows. Participants will leverage Claude Code integrated with Agent Skills, an AI coding assistant that can autonomously execute multi-step tasks, to analyze a large public-interest dataset provided by the challenge. The goal is to produce transparent AI processes that are inspectable, challengeable, correctable, and reusable, ensuring that the AI's reasoning can be scrutinized and refined by human journalists.

The challenge aims to address critical concerns in AI-augmented journalism: lack of transparency, difficulty in correcting errors, and the black-box nature of many AI models. By focusing on reusable workflows, Northwestern hopes to create a library of best practices that newsrooms can adopt for investigations ranging from public records analysis to pattern detection in large datasets. The use of Claude Code—an agentic coding tool that can write, edit, and execute code iteratively—enables participants to build solutions that are not only powerful but also auditable. This competition represents a significant step toward establishing ethical and practical standards for AI in journalism, potentially reshaping how newsrooms handle data-driven investigations.