🤖 AI

Practitioner Beliefs and Behaviors in AI-Enhanced Education: DOT Framework Survey Evidence

This study utilizes a cross-sectional survey of 72 higher education practitioners to validate the DOT Framework, revealing that while educators hold favorable views of AI as a pedagogical support tool and emphasize human oversight, their current implementation practices often lack systematic design elements and are hindered by institutional barriers such as limited policy and training.

Original authors: David Gibson (Curtin University), M. Elizabeth Azukas (Georgia Institute of Technology), Gerald Knezek (University of North Texas)

Published 2026-05-29

📖 4 min read☕ Coffee break read

CC BY 4.0

Original authors: David Gibson (Curtin University), M. Elizabeth Azukas (Georgia Institute of Technology), Gerald Knezek (University of North Texas)

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine a group of 72 teachers and educational experts gathered in a room to discuss a new, powerful tool they've all been experimenting with: Artificial Intelligence (AI). This paper is a report on what they think, how they use it, and what's stopping them from using it better.

Here is the story of that report, told in simple terms.

The Big Idea: The "DOT" Framework

The researchers used a specific map called the DOT Framework to understand these teachers. Think of this framework as a recipe for cooking with a very smart, but sometimes unreliable, sous-chef (the AI).

D (Design Thinking): This is the "cooking process." You don't just throw ingredients in a pot; you taste, adjust, and try again.
O (Open Systems): This is the "kitchen environment." It's not just about the chef and the sous-chef; it's about whether the kitchen has good ovens, if the manager has a rulebook, and if the customers (students) are happy.
T (Transformation): This is the final meal—the actual teaching that happens in the classroom.

The study asked: Are teachers following this recipe, or are they just winging it?

What the Teachers Actually Do (The "Winging It" Part)

The survey found that these teachers are very familiar with AI. In fact, most of them are already using it, but mostly for the "easy" parts of cooking.

What they love: They use AI to write lesson plans, create handouts, and brainstorm ideas. It's like using the sous-chef to chop vegetables or mix a batter.
What they avoid: They are very hesitant to let the AI grade papers or give final feedback to students. They treat the AI like a smart intern who can do the work, but never like the boss who makes the final decision.
The Missing Steps: The recipe says you should start by asking, "What does the student need?" (Needs Assessment) and end by asking, "Did this work?" (Feedback). The study found that teachers are great at the middle part (making stuff), but they often skip the beginning and the end. They are cooking without checking the recipe or tasting the final dish.

What They Believe (The Three Pillars)

When the researchers asked the teachers to rate their beliefs about AI, the answers grouped into three clear categories:

The "Super-Intern" Belief: Teachers think AI is great at doing tasks quickly. They believe it can help them be more creative and efficient.
The "Safety First" Belief: This was the strongest belief of all. Almost everyone agreed: "AI makes mistakes, so a human must always check its work." They believe AI should never replace a teacher's judgment. It's like trusting a GPS, but always keeping your eyes on the road.
The "Partner" Belief: Teachers see AI as a collaborator. They don't want a robot to take over; they want a partner to help them plan and organize.

The Roadblocks (The Kitchen is Broken)

Even though the teachers want to use this "sous-chef" effectively, they are hitting walls. The study found that the problem isn't the teachers; it's the kitchen itself.

No Rulebook: Most schools don't have clear policies on how to use AI.
No Training: Teachers are teaching themselves (self-directed learning) because the school isn't providing enough training.
No Tools: Sometimes the technology or infrastructure just isn't there.

The researchers describe this as a "macro-system bottleneck." Imagine a race car driver (the teacher) who is ready to drive fast, but the track (the school) has no guardrails, no pit crew, and no traffic lights. The driver is stuck.

The Bottom Line

This paper doesn't claim that AI is the future of education or that it will solve all problems. Instead, it offers a snapshot of the present:

Teachers are cautious but curious. They are using AI as a helper, not a replacement.
They are following part of the recipe. They are good at brainstorming and creating, but they are skipping the steps of checking student needs and evaluating results.
The system is holding them back. Without school-wide support, clear rules, and better training, teachers are left to figure it out alone.

The researchers suggest that for AI to truly help education, schools need to stop treating it like a shiny new gadget and start treating it like a complex system that needs a solid foundation, a clear rulebook, and a team effort to work properly.

Technical Summary: Practitioner Beliefs and Behaviors in AI-Enhanced Education (DOT Framework Survey Evidence)

Problem Statement
The rapid integration of generative artificial intelligence (AI) and large language models (LLMs) into higher education has outpaced the development of coherent, pedagogically grounded institutional strategies. While adoption is widespread, the absence of shared norms and frameworks has led to fragmented, ad hoc practices among educators. This fragmentation results in an instructional environment where the transformative potential of AI remains unrealized and risks remain unmanaged. The study addresses the lack of empirical data regarding how educators conceptualize AI's pedagogical role, their current usage behaviors, and the institutional conditions that facilitate or constrain responsible integration.

Methodology
The study employed a cross-sectional survey design to examine the beliefs, behaviors, and institutional conditions of 72 higher education practitioners in the United States. The sample, recruited from educational technology conferences and professional networks, was predominantly composed of full-time faculty (54%) and instructional designers (14%), with a high level of self-reported AI engagement (76% actively using or researching AI).

The instrument, the "DOT Survey," was developed based on the DOT Framework, a theoretical model integrating Design Thinking (D) and the Open Systems Model (O) to guide the Transformation (T) of teaching practices. The survey included:

19 Likert-scale belief items designed to capture views on AI's instructional potential, governance, automation capacity, and collaborative role.
Usage pattern questions across 11 task categories (e.g., content generation, grading, research).
Reflective practice questions assessing the frequency of design-thinking behaviors (needs assessment, outlining goals, experimentation, feedback collection).
Institutional challenge assessments regarding policy, training, and infrastructure.

Data analysis included descriptive statistics, reliability testing (Cronbach's alpha), and Exploratory Factor Analysis (EFA) using Principal Axis Factoring with Promax rotation to identify underlying constructs within the belief items.

Key Results

Factor Structure of Beliefs: EFA of the 19 belief items revealed a robust three-factor structure (Total Variance Explained: 54.4%, $\alpha$ = .90):
1. AI Functional Capabilities ( $\alpha$ = .910): Beliefs regarding AI's efficiency and effectiveness in specific tasks (feedback, grading, creativity, and serving as a teaching assistant).
2. Oversight and Governance ( $\alpha$ = .786): A principled commitment to human oversight, critical evaluation of AI outputs, transparency, and faculty control over materials. This factor showed high agreement (means > 4.60) but some ceiling effects.
3. Instructor Collaboration and Planning ( $\alpha$ = .851): Beliefs in AI as a collaborative partner that supports instructional planning, inclusivity, and instructor agency.
- Notable Finding: The item "AI can replace instructor expertise" failed to load on any factor and was nearly universally rejected (71% disagreement), reinforcing a "co-intelligence" rather than "replacement" mindset.
Usage Patterns: Practitioners reported high usage in generative and content-creation tasks (e.g., generating instructional materials, $M=3.43$ ; supporting research, $M=3.21$ ). Conversely, usage in evaluative and automated tasks was significantly lower (e.g., automating grading, $M=1.60$ ; providing personalized feedback, $M=2.17$ ).
Design Thinking Behaviors: While iterative prompting and experimentation were common, systematic adherence to the full design cycle was inconsistent. "Identifying needs before using AI" and "collecting feedback to iterate" were the least frequent practices, indicating a gap between the framework's theoretical requirements and current implementation.
Institutional Barriers: The primary challenges identified were ethical concerns/academic integrity (64%), insufficient training (61%), lack of institutional policies (50%), and limited infrastructure (47%). The data suggests a macro-system bottleneck where faculty are advancing faster than institutional support structures.

Key Contributions

Empirical Examination of the DOT Framework: The study provides preliminary descriptive support for the DOT Framework as a valid model for understanding practitioner beliefs, demonstrating that educator perspectives cluster in ways consistent with the framework's design thinking and open systems architecture.
Measurement Instrument Development: It introduces and evaluates a survey instrument capable of capturing design-oriented and systems-aware dimensions of AI use, offering a three-factor structure (Functional Capabilities, Oversight/Governance, Collaboration/Planning) for future research.
Identification of Implementation Gaps: The study highlights specific discrepancies between the full design cycle proposed by the DOT Framework and current practitioner behaviors, specifically the underutilization of the "Empathize" (needs assessment) and "Test/Feedback" stages.

Significance and Claims
The authors position this study as exploratory and preliminary, explicitly stating that it does not establish causal relationships between beliefs, behaviors, and instructional outcomes. The significance lies in:

Descriptive Validation: Providing initial evidence that the DOT Framework's conceptual distinctions (micro-system design vs. macro-system governance) map onto actual practitioner cognition.
Highlighting the Macro-System Bottleneck: Confirming that sustainable AI integration is constrained not by individual faculty capability but by institutional deficits in policy, training, and infrastructure.
Defining the "Co-Intelligence" Norm: Demonstrating that engaged practitioners largely reject AI as a replacement for expertise, instead viewing it through a lens of critical collaboration and human oversight.

The paper concludes that future research must move from descriptive modeling to confirmatory validation (using CFA) and outcome-based studies to determine how DOT-aligned practices influence instructional quality. It calls for institutions to treat professional development and governance structures not as optional enhancements, but as prerequisites for ethical and effective AI integration.

The Big Idea: The "DOT" Framework

What the Teachers Actually Do (The "Winging It" Part)

What They Believe (The Three Pillars)

The Roadblocks (The Kitchen is Broken)

The Bottom Line

More like this