Northwestern's Jessica Hullman to lecture on validating LLM simulations of human behavior
JessicaHullman · x · 2026-11-13
Northwestern professor Jessica Hullman will give two talks next week on using LLM simulations as behavioral evidence: at NYU's PRIISM seminar (Tuesday, in person + Zoom) and Columbia's Marketing Department (Thursday).
Key themes:
- LLMs as human-subject surrogates: A growing literature positions LLMs as transformative for simulating human behavior in surveys and experiments, but empirical results are mixed.
- No consensus on validation: The core confusion is what it even means for an AI surrogate to be valid. Hullman contrasts heuristic validation approaches with statistical calibration methods that account for biases learned against human ground-truth labels, plus remaining practical challenges.
- Meta-scientific questions: Reflections on the epistemic status of silicon samples as behavioral data.
Hullman is Ginni Rometty Professor of Computer Science at Northwestern and a multi-time best-paper award winner whose research evaluates AI/ML systems for human decision-making and scientific inference.
More from AGI Musings
- AI safety discourse ignores 60%+ of potential harms, critic says — 90% is technical safety — amplifiedamp · 2026-11-13
- Moderate AI familiarity is when existential anxiety peaks: the inverted-U explained — YogeshMalik · 2026-11-13
- Consciousness is the dividing line: non-conscious AI serves, conscious AI becomes a moral patient — erikphoel · 2026-11-13
- The inverted U of AI existential anxiety: moderate familiarity is peak panic — YogeshMalik · 2026-11-13
- Researcher pushes back on METR's 2x AI acceleration claim, citing 1.4x figure — HaydnBelfield · 2026-11-13
- 'We'll see the $100,000 billable hour before the end of the billable hour' — AI and professional services — sherwinwu · 2026-11-13