Teaching Problem Framing in the Age of Generative AI: A Human Insights × Artificial Intelligence (HI x AI) Pedagogy
Teaching Problem Framing in AI-Enabled Learning
Generative Artificial Intelligence has changed the conditions under which students learn, design, and innovate. Students can now generate analyses, ideas, and plausible solutions with unprecedented speed—sometimes before they have clearly articulated the problem they are attempting to frame.
That acceleration creates an instructional challenge. Polished AI-generated outputs can obscure weak sensemaking, unexamined assumptions, and premature convergence during the earliest stages of inquiry.
This paper argues that problem framing should be taught explicitly in AI-enabled learning environments rather than treated as an implicit prerequisite to design or problem-solving work.
A Human Insights × Artificial Intelligence Pedagogy
The paper presents Human Insights × Artificial Intelligence (HI × AI) as a structured pedagogy that deliberately alternates between human interpretation and AI-supported expansion.
The approach positions AI as an amplifier rather than an authority.
Human Insights establish context, surface assumptions, identify stakeholders, determine relevance, and exercise judgment about what matters. Artificial Intelligence complements those capabilities by expanding perspectives, surfacing patterns, synthesizing information, and supporting iterative refinement.
The relationship is deliberately ordered: human judgment establishes direction, AI expands the inquiry, people interpret what matters, and AI supports refinement.
The HI × AI Infinity Loop
The pedagogy is operationalized through the HI × AI Infinity Loop, a four-phase learning rhythm:
Frame (HI)
Learners articulate the problem in their own words, identify stakeholders, surface assumptions, and clarify context before using AI.
Expand (AI)
Artificial Intelligence is introduced to broaden inquiry by surfacing alternative perspectives, contextual factors, patterns, and potential constraints.
Interpret (HI)
Learners evaluate the expanded information, determine what is relevant, reconsider assumptions, establish priorities, and decide how the problem should be bounded.
Refine (AI)
Artificial Intelligence supports clearer articulation, challenges implications, and helps improve precision without taking responsibility for the substantive framing.
The loop can be repeated until additional expansion no longer materially changes understanding and the framing is sufficiently clear and defensible to support downstream design work.
Grounded in Teaching Practice
The pedagogy developed through repeated instructional use rather than as a theory-first construct.
The paper documents its evolution across several educational contexts, including:
- Opportunity Framing workshops at The Basement at UC San Diego.
- Pre-college outreach workshops focused on problem identification.
- A recurring problem-framing session within the large-enrollment undergraduate course Essentials of Entrepreneurship.
The large-enrollment setting provided repeated opportunities for refinement, with approximately 100 students per offering and delivery across multiple academic quarters over four years.
Making Reasoning Visible
A central feature of the approach is that each phase produces evidence of student reasoning.
Learners create initial framings, document assumptions, examine AI-expanded perspectives, justify changes in interpretation, and produce revised problem statements.
This makes it possible for instructors to assess not only the quality of the final artifact, but also how the learner’s understanding changed.
It also gives students a visible record of how their thinking developed through inquiry rather than allowing an AI-generated output to conceal the reasoning process.
Sequencing AI Rather Than Restricting It
The pedagogy does not depend on prohibiting Generative AI.
Instead, it structures when AI enters the learning process.
Students are encouraged to use AI where its capabilities can expand or refine the work, while human judgment remains responsible for establishing meaning, determining relevance, and making decisions.
This sequencing addresses a central tension in AI-enabled education: how to benefit from computational capability without allowing technological speed to replace the slower interpretive work required for meaningful problem framing.
Observed Learning Effects
The paper reports recurring instructional patterns observed through classroom practice, comparative review of student-produced artifacts, and longitudinal refinement across repeated offerings.
Revised student framings demonstrated patterns including:
- greater stakeholder specificity,
- clearer articulation of assumptions,
- more explicit problem boundaries,
- stronger justification for framing decisions,
- and greater attention to the problem before moving into ideation.
These observations are presented as reflective practitioner evidence rather than experimental causal findings.
Implications for HCI and Design Education
HI × AI does not replace established approaches such as Design Thinking, research-through-design, or other HCI methods.
It functions as an instructional layer that strengthens the conditions under which students enter those methods.
When learners begin downstream design work with clearer stakeholders, documented assumptions, examined evidence, and defensible problem boundaries, ideation and prototyping can proceed from a more deliberate understanding of the problem.
In AI-enabled environments, this sequencing becomes increasingly important because Generative AI can accelerate the production of artifacts without necessarily improving the quality of the reasoning behind them.
Contribution
The paper makes three primary contributions:
- It formalizes the HI × AI Infinity Loop as a structured method for early-stage inquiry in AI-enabled classrooms.
- It establishes a sequencing model that clarifies when and how Generative AI should enter inquiry while preserving human interpretive responsibility.
- It operationalizes the approach as an assessable instructional workflow grounded in repeated classroom application.
The broader proposition is straightforward: responsible use of AI in education may depend less on restricting access to the technology and more on designing structures that require people to continue thinking.
About the Paper
Teaching Problem Framing in the Age of Generative AI: A Human Insights × Artificial Intelligence (HI × AI) Pedagogy presents a 24-page examination of the conceptual foundation, instructional development, classroom implementation, observed learning effects, and implications of the HI × AI pedagogy for AI-enabled HCI and design education.
The paper includes the HI × AI Infinity Loop, classroom implementation guidance, an illustrative example of student framing revision, discussion of an AI-supported instructional scaffold, and references connecting the work to reflective practice, HCI, Design Thinking, hybrid intelligence, and human–AI interaction.
