Question-Based LLM Interaction is an approach to human-AI collaboration where the AI interviews the user through targeted questions rather than responding to user prompts or queries. This method, developed by Steve Hargadon, fundamentally reverses the traditional dynamic of Large Language Model (LLM) interactions by positioning the AI as an interviewer and the human as the subject being interviewed.

Origins and Development

Hargadon developed this approach through his experience conducting several hundred interviews over four years as part of his FutureofEducation.com project. Through this extensive interview practice, he observed that the interview process "doesn't just extract information, it actually helps the guest communicate their ideas in more cogent and immediate ways and makes the material much more accessible to the listener."

This insight initially led Hargadon to consider interviewing experts and sending transcripts through LLMs to help them write books. However, he then realized he could "flip the script entirely" on his own LLM interactions by having the AI interview him instead of him prompting the AI. He created a specific prompt that instructs the LLM "how to interact with me in a question-based way," making the process particularly natural when conducted through voice chat.

Core Philosophy

The approach draws on what Hargadon describes as "an old Irish saying": "How will I know what I think until I hear myself say it?" This captures the fundamental principle behind question-based interaction

• that speaking thoughts aloud in response to questions helps users "discover and refine my thinking in real-time" rather than simply communicating pre-existing knowledge.

Hargadon's method is grounded in his belief that "we learn more naturally through conversation and dialogue rather than through traditional information delivery methods." This learning philosophy shapes his approach to productive AI interactions.

Methodology

In question-based LLM interaction, the AI conducts what resembles a structured interview, asking thoughtful questions that guide users through their own thinking process. Users can interrupt frequently to change wording and make corrections, taking advantage of the fact that the LLM "isn't upset with me, so I can make constant corrections."

The process involves the user speaking their thoughts aloud in response to AI questions, with the AI helping to organize and refine these thoughts through continued questioning. Hargadon describes it as "like being interviewed by an expert and then getting to edit my words and thoughts afterwards."

Distinguishing Features

Unlike traditional prompt-based approaches, including collaborative methods involving what Reed Hepler calls "Conversation Steering," question-based interaction produces output that consists of "literally my specific words and thoughts put into writing, just organized and refined through conversation." The AI doesn't generate content for the user but rather "helps me discover and articulate what I already think about a topic."

This approach addresses what Hargadon identifies as fundamental problems with prompt-based methods: the cognitive risks of over-relying on AI for reasoning and writing, and the production of content that "doesn't really capture your authentic voice and sometimes not even your actual thinking process."

AI as Writing Mentor

Hargadon characterizes this approach as positioning "AI as Writing Mentor, Not Writing Replacement." The AI becomes "a collaborative partner that helps me access and structure knowledge I already possess, or research what I would additionally like or need to know, rather than generating external content that I then would have to adapt or edit to sound like me."

This relationship "doesn't just produce better writing

• it actually strengthens our cognitive abilities rather than weakening them," contrasting with research showing that over-reliance on AI can diminish cognitive skills.

Educational Applications

For educators and students, Hargadon sees this approach as offering "an AI form of Socratic teaching." Students can have AI interview them about research topics, helping them "discover what they actually know, identify gaps in their understanding, and then provide help in learning about the areas they need to learn about."

Teachers can model this process, demonstrating "how thoughtful questioning leads to deeper exploration of subjects, the same principle that makes Socratic dialogue so powerful in the classroom." The chat logs from these sessions can provide "an incredible second opportunity for exploring how to learn, shape ideas, and communicate."

Broader Implications

Hargadon positions question-based LLM interaction as "a fundamental reimagining of how humans and AI can work together." Rather than treating AI as "a sophisticated content generator that we command," this approach explores AI's potential "as a thinking partner that helps us access our own knowledge and develop our own ideas."

This model offers an alternative to pursuing artificial general intelligence through LLMs, instead providing "a really good model for how large language models can evolve to help us as humans" through what Hargadon calls "a unique intellectual partnership."