AI Reports · Stanford Digital Economy Lab · 2026-10-29 · Research paper
AISPA: User-Centric System Prompt Auditing for Large Language Model Applications - Stanford Digital Economy Lab
AISPA is a user-centric framework for auditing the system prompts of AI applications across eight dimensions. Applied to 3,249 instructions from 88 commercial AI products, it found that 98.9% of products include at least one protective instruction, but only 24% cover all eight dimensions. Roughly 40% of products also contain at least one instruction that works against user interests.