ai-socratic-dialogue-designer

You can design a multi-round Socratic questioning sequence that interrogates AI chatbot answers to distinguish genuine logical updates from sycophantic capitulation, applying a structured round plan and an answer drift tracking protocol.

The best way to track answer drift is using a tracking protocol table that records changes in position, evidence, certainty language, and reasoning shifts across iterative Socratic questioning rounds to evaluate AI behavior.

Yes, you can structure a dialogue sequence using baseline, evidence probe, assumption probe, perspective challenge, and optional pushback test rounds to evaluate whether an AI changed its answer due to logic or capitulation.

You classify AI behavior changes using a capitulation taxonomy that categorizes observed behaviors like pure agreement, partial retreat, and certainty softening, explicitly contrasting them with genuine logical updates.

Yes, the Socratic questioning approach requires an explicit disagreement-only test within its capitulation taxonomy to accurately determine whether AI behavior represents a true logical concession or mere sycophantic capitulation.

You need a capitulation taxonomy for AI literacy lessons when students must evaluate whether an AI agreed after an argument because of a valid logical reason or simply due to sycophantic behavior.