AI Can't Discern Whether a Topic Is Problem-Driven or Just Interesting

A student sent me her “dissertation topic” last week: “Leadership in healthcare organizations during times of change.” “ChatGPT helped me develop this,” she said. “It’s focused on an important area.” I asked: “Which aspects of leadership? Which healthcare organizations? What kinds of change? What specifically will you study?rdquo; She paused. “Well… just leadership generally during any organizational change?rdquo; That’s not a topic. That’s an entire library section. Her committee would reject it immediately for being impossibly broad. But AI had convinced her it was sufficiently focused because it mentioned a field (healthcare), a concept (leadership), and a context (change). Here’s what students don’t understand: AI generates broad themes that sound focused but aren’t. Doctoral success requires the systematic narrowing process—asking refinement questions repeatedly until you reach the precision committees require. AI can’t do this iterative refinement because it doesn’t understand how to move from general interest to specific defensible research question.


Good Topics Emerge Through Iterative Refinement


Let me show you how topic development actually works versus what AI produces.

What AI Produces: Single-Pass Broad Suggestions


You tell AI: “I’m interested in teacher motivation” AI responds: “Consider studying factors affecting teacher motivation in schools, examining how various organizational and individual variables influence teacher motivation and its relationship to student outcomes.” This sounds like topic development, but it’s just your broad interest rephrased with academic language. AI hasn’t narrowed anything—it’s made your vague interest sound scholarly while keeping it equally vague.

What Real Development Looks Like: Multiple Refinement Passes


Starting point: “I’m interested in teacher motivation” Pass 1 – Specify which aspect: “What about motivation? Its causes? Its effects? How it changes over time?” Refined: “What causes teachers to stay motivated” Pass 2 – Identify specific factors: “Which causes? Organizational? Personal? Leadership-related?” Refined: “How organizational support affects teacher motivation” Pass 3 – Specify population: “Which teachers? All teachers? Specific grade levels? Specific contexts?” Refined: “How organizational support affects motivation among early-career elementary teachers” Pass 4 – Add context: “In what settings? High-performing schools? Struggling schools? Urban? Rural?” Refined: “How organizational support affects motivation among early-career elementary teachers in high-poverty urban schools” Pass 5 – Specify mechanism or moderator: “Does anything make this relationship stronger or weaker?” Refined: “How organizational support affects motivation among early-career elementary teachers in high-poverty urban schools, and whether this relationship is moderated by teacher self-efficacy” Now you have a precise, defensible topic. This took five refinement passes—something AI never does. According to research from Harvard Graduate School of Education, students who engage in systematic topic refinement through multiple iterative cycles complete dissertations 18 months faster on average than students who attempt to proceed with insufficiently refined topics.


What AI Outputs Instead of Precision


AI generates content that seems focused but lacks the specificity doctoral work requires.

Buzzwords Without Substance


AI uses impressive-sounding academic terminology without adding precision: AI output: “Examining the intersection of digital transformation and organizational culture in the context of leadership development” Problem: This strings together buzzwords (digital transformation, organizational culture, leadership development) without specifying what you’re actually studying. What aspect of digital transformation? Which dimensions of organizational culture? What about leadership development specifically? Precise version: “How healthcare administrators describe organizational culture changes during electronic health record implementation and whether these perceived changes predict participation in voluntary leadership training programs” Notice how precision replaces buzzwords with specific, measurable concepts.

Generic Phrases That Sound Academic


AI loves phrases that appear scholarly but communicate nothing specific: AI favorites:
  • “Factors affecting…”
  • “Various aspects of…”
  • “Key dimensions of…”
  • “Important considerations in…”
  • “Critical elements of…”
Problem: These phrases are placeholders for actual specificity. “Factors affecting teacher retention” could mean hundreds of different studies depending on which factors, which teachers, and which contexts. Precise alternative: “The relationship between principal supportive behaviors (as measured by the Supportive Principal Leadership Scale) and turnover intention among secondary teachers in rural districts, with years of experience as a moderator”

Popular Themes Without Novel Angles


AI suggests topics that are currently popular in academic discourse without identifying what would make your study original: AI suggestions:
  • “Artificial intelligence in education”
  • “Remote work and organizational culture”
  • “Healthcare equity and access”
  • “Climate change and organizational sustainability”
These are hot topics, but what specifically about them will you study? What hasn’t been studied? What’s your unique angle? AI doesn’t answer these questions.


The Human Narrowing Process


Let me walk you through exactly how experienced advisors guide systematic narrowing—the process AI cannot replicate.

Step 1: Identify Your Core Interest


Start very broad: “I’m interested in employee turnover” This isn’t a topic yet, but it’s a legitimate starting point. We need to narrow through systematic questioning.

Step 2: Add Variable Specification


Question: What about turnover? Its predictors? Its consequences? Specific types of turnover (voluntary vs. involuntary)? Decision processes? Refinement: “I want to understand what predicts voluntary turnover” Next question: Which predictors? Job satisfaction? Leadership? Compensation? Work-life balance? Organizational culture? Refinement: “I want to understand how supervisor support predicts voluntary turnover” Now we have an independent variable (supervisor support) and dependent variable (voluntary turnover).

Step 3: Refine Sample Demographics


Question: Which employees in which organizations? Too broad: “All employees” More specific: “Healthcare workers” Even more specific: “Registered nurses in hospitals” Most specific: “Registered nurses in intensive care units at urban teaching hospitals” Each refinement layer increases feasibility and potential for originality.

Step 4: Add Moderating or Mediating Variables


Question: What else might matter? What might make supervisor support more or less predictive of turnover? Potential moderators:
  • Employee burnout levels (maybe support matters more when burnout is high)
  • Hospital resources (maybe support is only effective when hospitals have adequate resources)
  • Nurse experience (maybe new nurses benefit from support differently than experienced nurses)
Refinement: “How supervisor support predicts voluntary turnover intention among ICU nurses at urban teaching hospitals, with burnout level as a moderator” Adding complexity creates sophistication without sacrificing clarity.

Step 5: Adjust Methodology Type


Question: How should this be studied? What’s missing from existing research methodologically? If existing research is all quantitative surveys: “Qualitative interviews exploring how ICU nurses describe supervisor support and how it influences their turnover decisions during high-burnout periods” If existing research is all qualitative: “Quantitative survey testing whether supervisor support (measured by Leader-Member Exchange scale) predicts turnover intention (measured by Turnover Intention Scale) among ICU nurses, with burnout (measured by Maslach Burnout Inventory) as a moderator” If no recent research exists: “Mixed methods study examining supervisor support and turnover among ICU nurses: quantitative survey of relationships followed by qualitative interviews exploring mechanisms” Methodological specification creates both precision and originality.

Step 6: Identify Under-Studied Contexts


Question: Where has this not been studied? What contexts create unique dynamics? Understudied contexts might include:
  • During staffing crises or pandemics
  • In hospitals serving rural vs. urban populations
  • In safety-net hospitals with different resource constraints
  • During leadership transitions or organizational restructuring
Final refined topic: “The moderating role of burnout in the relationship between supervisor support and turnover intention among ICU nurses in urban safety-net hospitals during the post-pandemic staffing crisis: A cross-sectional survey study” This is defensibly narrow, addresses a specific problem, includes novel contextual elements, and is clearly researchable.


Why Precision Equals Approval and Speed


Insufficiently narrowed topics create specific problems that delay graduation. Let me show you exactly how precision accelerates completion.

Problem 1: Unfocused Literature Reviews


Broad topic: “Leadership and organizational culture” Literature review problem: You need to review everything about leadership (hundreds of theories, thousands of studies) and everything about organizational culture (equally vast). Your Chapter 2 balloons to 100+ pages and still feels incomplete. Your committee says it’s too broad. Precise topic: “The relationship between nurse managers’ transformational leadership behaviors and unit-level safety culture in trauma centers” Literature review benefit: You review transformational leadership specifically (not all leadership), safety culture specifically (not all culture), in healthcare contexts (not all organizations). Focused 40-page review that thoroughly covers relevant research.

Problem 2: Unclear Research Questions


Broad topic: “Technology integration in education” Research question problem: “How does technology integration affect education?” is too vague to guide research. What technology? What aspects of education? Measured how? Precise topic: “The relationship between teacher self-efficacy for technology integration and frequency of using adaptive learning software in middle school math instruction” Research question benefit: “To what extent does teacher self-efficacy for technology integration predict frequency of adaptive learning software use in middle school math classrooms?” is specific and testable.

Problem 3: Overwhelming Data Collection


Broad topic: “Factors affecting student achievement” Data collection problem: To study this comprehensively, you’d need data on dozens of factors (family background, prior achievement, teacher quality, school resources, peer effects, instructional quality, curriculum, etc.). Impossible scope. Precise topic: “The relationship between teacher feedback specificity and student self-efficacy in 9th grade English classes, with prior achievement as a covariate” Data collection benefit: Focused measures of specific constructs. Manageable survey or observational protocol addressing exactly what you need.

Problem 4: Prolonged Approval Process


Broad topic: “Social media and adolescent development” Approval problem: Committee keeps asking for more specificity. Multiple revision rounds trying to narrow scope. Months wasted in revision loops. Precise topic: “The relationship between Instagram use frequency and body image concerns among 13-15 year old girls in suburban middle schools, with peer social comparison tendency as a mediator” Approval benefit: Committee can immediately assess whether this is viable, original, and appropriately scoped. Quick approval or specific feedback for minor adjustments.

The Time Savings


Students with insufficiently narrow topics:
  • 6-12 months: Revising proposals trying to narrow scope
  • 6-12 months: Collecting data for overly broad studies
  • 6-12 months: Analyzing and interpreting overwhelming amounts of data
  • Total: 18-36+ months from proposal approval to defense
Students with properly narrow topics:
  • 2-3 months: Minor proposal revisions (topic already well-defined)
  • 3-6 months: Focused data collection on specific constructs
  • 3-6 months: Clear analysis addressing specific research questions
  • Total: 8-15 months from proposal approval to defense
Precision saves a year or more of your doctoral program.


Get Expert Help With Systematic Narrowing


Don’t let AI’s inability to narrow topics trap you in years of unfocused work. Get guidance from advisors who understand the iterative refinement process.

Our Topic Refinement Process


We guide you through systematic narrowing: Session 1: Identifying core interests What genuinely interests you? What problems do you care about? What populations or contexts? Session 2: Variable specification Which specific variables, constructs, or phenomena? What relationships or processes? Session 3: Population and context refinement Which specific populations? Which contexts create unique dynamics or gaps? Session 4: Methodological positioning What methods address your questions? What methodological gaps exist? Session 5: Verification and positioning Is this narrow enough? Too narrow? How do we verify originality? How do we position for committee approval? Get systematic topic refinement help that moves you from broad interest to precise, defensible topic.

From Broad to Specific: Documented Process


We don’t just give you a narrow topic—we teach you the refinement process: Documentation of narrowing: We show you exactly how your topic evolved from broad to specific, so you can explain your logic to committees Alternative refinements: We explore multiple narrowing paths, helping you understand trade-offs between different specific directions Originality verification: At each refinement stage, we search literature to assess whether you’ve reached sufficient specificity for originality Feasibility checking: Ensuring your refined topic remains feasible while achieving necessary precision

Complete Dissertation Support


Topic refinement is the foundation, but we provide ongoing support: Get comprehensive dissertation help ensuring your refined topic translates into strong proposals, focused data collection, and timely completion.


The Bottom Line: AI Produces Themes, Humans Create Topics


AI generates broad academic themes that sound focused but lack the precision doctoral work requires. It cannot perform the iterative refinement process that creates defensible research topics. Only human advisors can:
  • Ask refinement questions systematically across multiple passes
  • Recognize when topics remain too broad versus appropriately narrow
  • Balance precision with significance (not so narrow it’s trivial)
  • Verify that refined topics are original yet feasible
  • Predict whether committees will view topics as appropriately scoped
Don’t attempt to proceed with AI-generated broad themes masquerading as topics. Work with experts who understand that good topics emerge through systematic narrowing, not through single-pass AI generation. The difference between finishing in 3 years versus 5+ years often comes down to starting with properly narrowed topics versus attempting to proceed with insufficiently refined ideas.
Scroll to Top