Ancient Socratic Questioning Methods Mirror Modern AI Prompt Engineering Techniques

AI researcher Li Fei-Fei sparked discussion by suggesting Socrates would have been the world's best prompt engineer, given his method of using structured questions to expose contradictions and refine thinking. Philosopher Socrates (470–399 BC) never provided direct answers in his dialogues but instead used layered questioning to guide others toward clearer reasoning — a technique closely paralleling how effective LLM prompts are designed today. Scholar R.W. Paul identified six core Socratic question types — clarifying concepts, exposing assumptions, requesting evidence, shifting perspective, tracing consequences, and questioning the question — each of which maps to a recognized modern prompting technique such as chain-of-thought or adversarial checking. Just as Socrates used 'elenchus' to systematically dismantle flawed arguments made by Sophists, prompt engineers today use counterexample-based instructions to reduce fluent but inaccurate AI outputs. The article argues that rather than being new innovations, advanced prompting strategies are largely modern repackagings of methods Socrates practiced over 2,400 years ago.