Three Recent AI Encounters That Prove There's No Going Back
The Adaptive Integration Approach: Finding Balance in AI Education
If you or your institution find value in the pedagogical and instructional explorations at Educating AI, consider becoming a paid subscriber.
Your support helps sustain independent research rooted in critical analysis, practical insight, and a commitment to thoughtful innovation.
Introduction
As artificial intelligence rapidly transforms the educational landscape, institutions find themselves at a crossroads. On one side stands the protectionist mindset—focused on detecting and curbing AI use to preserve traditional assessment integrity. On the other, the expansionist vision—demanding a complete curricular overhaul to focus exclusively on AI upskilling. Both approaches, in their extremes, fail to adequately prepare students for a future where human expertise and artificial intelligence increasingly intertwine.
In this article, I propose a third way: the Adaptive Integration Approach. This balanced perspective seeks AI-responsive systems that maintain academic integrity while thoughtfully incorporating AI tools into educational processes. It preserves instruction in core competencies that foster higher-order thinking while embracing the transformative potential of AI for 21st-century professional development.
Three recent personal experiences with advanced AI technology have convinced me that neither pure detection nor wholesale expansion adequately addresses the profound shifts occurring in education. Let me share these experiences to illustrate why a more nuanced approach is necessary.
Experience 1: Collaborative Theoretical Development with Advanced AI
My recent work with ChatGPT o3 to develop an extension of rhetorical narratology illustrates the remarkable collaborative potential of advanced AI. I sought to incorporate the generative fiction of AI models into Phelan's rhetorical communication model of author, resources, and audience. While not entirely successful, the experience revealed something profound about emerging AI capabilities.
The model's ability to seamlessly riff on the flow of my thoughts as they developed throughout the writing and research process was truly outstanding. When text generation failed, the failures themselves sparked new ideas. When it produced breakthrough concepts, I became convinced I was experiencing a new level of AI capacity. Yes, it remained clunky at times. Yes, it failed when overclustered with ideas. Yes, it fabricated quotes. But something genuinely interesting is churning in those circuits—something that signals significant changes on the horizon.
Looking back at models from just two years ago makes the acceleration of progress startlingly clear. Our students will soon have free access to tools that deeply simulate focused analytical processes. How could we possibly detect this out of existence? Much of student reliance on this kind of machine cognition could remain invisible beneath the surface of text production. The protectionist approach alone cannot address this reality.
Experience 2: Automated Assessment with Manus AI
My second experience involved experimenting with Manus AI to complete a low-stakes online course. Despite initial hesitation about giving control to the software, I utilized it within this limited scope to work through early sections of the course.
The results were eye-opening. Any online course relying on multiple-choice, short-answer, or true/false assessment formats is fundamentally vulnerable. Manus scans each page, analyzes content, works through question options, and answers correctly about 90% of the time. While currently slow and costly (I exhausted my credits within an hour), the technology represents a total game-changer for traditional assessment methods.
This raises crucial questions: How will our assessments evolve in response? Online individualized pathway education has become a major revenue stream for many institutions, yet AI inadvertently pushes us back toward classroom-based, human-centered assessment. The alternatives are limited: either return to in-person evaluation or embrace the reality that students will use AI in online settings and build pedagogical approaches on that foundation.
Neither pure detection nor pure expansion adequately addresses this challenge. We need thoughtful integration strategies that acknowledge these new realities while preserving meaningful assessment.
Experience 3: Specialized AI Research Tools Reshaping Information Access
My third experience involves exploring various specialized AI Deep Search options including those offered by Perplexity, ChatGPT, and Gemini. What I've discovered is a rapidly evolving ecosystem of research assistants, each with distinctive strengths that complement traditional research methods in different ways.
I now use different engines for different kinds of research tasks. Gemini excels at gathering numerous sources, creating a broad research foundation. ChatGPT produces the most well-written summaries, skillfully synthesizing complex information into coherent narratives. Perplexity offers live, adaptable research that responds dynamically to evolving queries.
This specialized differentiation signals another profound shift. Our students already have access to these powerful, increasingly nuanced research tools. The research process—traditionally a cornerstone of academic skill development—is being fundamentally transformed. Students armed with these tools can rapidly gather, synthesize, and reformulate information that would have previously required weeks of library work.
Detection approaches cannot meaningfully address this reality. How do we distinguish between a student who used Perplexity to gather sources versus traditional database searches? How do we detect when a student used ChatGPT to help synthesize research findings versus doing it entirely independently? More importantly, should we even try to make these distinctions when the professional world increasingly values the ability to effectively direct and leverage these AI research tools?
The research process itself is being reimagined before our eyes. Neither clinging to traditional research pedagogies nor abandoning the teaching of research principles serves our students well. We need an approach that acknowledges this new reality while ensuring students develop the critical thinking skills to evaluate, contextualize, and thoughtfully build upon AI-assisted research.
The Adaptive Integration Approach: A Third Way Forward
These three experiences illustrate why we need an approach that transcends the protectionist-expansionist dichotomy. The detection-focused protectionist mindset fights an increasingly unwinnable battle against rapidly advancing technology. The expansionist vision risks abandoning crucial foundational skills in its rush to embrace AI tools.
I propose the Adaptive Integration Approach as a balanced third way forward. This framework acknowledges the transformative impact of AI while maintaining core educational values and objectives. It adapts assessment and pedagogy to work with AI rather than against it, while preserving the development of fundamental human capabilities that technology cannot replicate.
This approach embodies several key principles:
Integrity Through Transparency: Rather than relying solely on detection, establish clear disclosure policies that acknowledge AI use while maintaining academic standards. This creates a culture of honesty rather than a technological arms race.
Contextual Assessment Design: Develop assessments that evaluate both AI-assisted and independent performance, making thoughtful distinctions about where each is appropriate based on learning objectives.
Critical AI Literacy: Teach students to be critical consumers and directors of AI tools—understanding their limitations, biases, and appropriate applications rather than simply using them as black-box solutions.
Core Competency Preservation: Maintain instruction in fundamental skills and knowledge that inform higher-order thinking, even as we incorporate AI tools that automate certain aspects of information processing.
Ethical AI Integration: Develop institutional guidelines that promote responsible AI use aligned with educational values and ethical principles.
Implementation Strategies: Putting Adaptive Integration into Practice
How do we implement this balanced approach in real educational contexts? Here are concrete strategies that institutions can adopt:
1. Tiered Assessment Framework
Develop a three-tiered assessment framework that clearly distinguishes between:
AI-Free Zones: Assessments that measure fundamental skills without AI assistance, conducted in controlled environments when necessary (e.g., in-class writing, proctored examinations).
AI-Assisted Zones: Assignments that explicitly permit and incorporate AI tools, with clear disclosure requirements and evaluation criteria focused on effective tool direction and critical evaluation of outputs.
AI-Collaborative Zones: Complex projects where students work iteratively with AI, documenting their process and demonstrating how they leveraged AI capabilities while applying human judgment and creativity.
This framework maintains spaces for developing independent capabilities while acknowledging AI's role in contemporary work processes.
2. Process Documentation Over Detection
Rather than investing heavily in increasingly fallible detection technologies, shift resources toward process documentation methods:
Require annotated drafts that show thinking evolution
Implement staged submission requirements that capture development phases
Use reflection components that articulate decision-making processes
Create AI disclosure statements that specify tool usage and contribution
These approaches focus on understanding how students develop ideas rather than policing the final product.
3. Skill Rebalancing in Curriculum
Rebalance curriculum emphasis away from skills that AI readily replicates toward higher-order capabilities:
Decrease emphasis on information gathering and increase focus on critical evaluation of sources
Reduce time spent on basic summarization and increase attention to nuanced analysis
Shift from memorization-heavy assessment toward application and extension
Emphasize creative problem-solving over procedural mastery
This rebalancing preserves what remains distinctly human while acknowledging areas where AI augmentation is becoming standard.
4. AI Literacy Across the Curriculum
Integrate AI literacy throughout disciplinary instruction rather than treating it as a separate subject:
Train faculty across departments in AI capabilities relevant to their field
Develop discipline-specific guidelines for appropriate AI use
Create assignments that explicitly teach effective prompting within subject contexts
Build critical analysis of AI outputs into regular coursework
This integration helps students understand AI as a contextual tool rather than a universal solution.
5. Institutional Policy Development
Develop comprehensive institutional policies that address:
Clear guidelines on permissible AI use by context and purpose
Transparent disclosure requirements that normalize honest reporting
Updated academic integrity frameworks that account for AI assistance
Faculty development resources for adapting to AI-augmented education
Assessment redesign support that maintains rigor while acknowledging new realities
These policies provide consistent frameworks while allowing disciplinary flexibility.
Case Study: Adapting a Research-Based Course
To illustrate the Adaptive Integration Approach in practice, consider how it might transform a traditional research-based course:
In the past, students would progress from topic selection through preliminary research, thesis development, source gathering, outlining, drafting, and revision. The instructor would assess final papers primarily on content, argumentation, and writing quality.
Under an Adaptive Integration model, the course might be redesigned as follows:
AI-Free Zone: In-class writing exercises to develop thesis statements and analytical frameworks without AI assistance, ensuring students can independently generate ideas.
AI-Assisted Zone: Source gathering phase where students explicitly use research tools like Perplexity or Gemini, but must provide critical evaluations of AI-suggested sources and demonstrate ability to refine search parameters effectively.
AI-Collaborative Zone: Drafting process where students may use ChatGPT for outlining or language refinement, but must submit annotated drafts showing their prompting strategies, editing decisions, and critical evaluation of AI-generated content.
Assessment Focus: Evaluation shifts from the final product alone to include process documentation, prompt development skill, critical engagement with AI outputs, and the student's unique analytical contribution. The final grade considers both independent thinking capability and effective AI direction.
Conclusion: Embracing the Inevitable, Preserving the Essential
The experiences I've shared with ChatGPT, Manus AI, and specialized research tools make one thing abundantly clear: AI is not merely a passing educational challenge to be detected and prevented, nor is it a replacement for developing core human capabilities.
The Adaptive Integration Approach offers a path forward that acknowledges the transformative—and inevitable—impact of AI on education while preserving what remains essential about the learning process. It replaces the false choice between detection and wholesale adoption with a nuanced framework that prepares students for a world where human and artificial intelligence increasingly collaborate.
By embracing transparency over detection, process over product, and critical engagement over either rejection or uncritical acceptance, we can develop educational approaches that remain relevant, rigorous, and responsive to the complex realities of our technology-augmented future.
The question before us is not whether AI will transform education—it already has. The question is whether we will shape that transformation thoughtfully or allow it to shape us by default. The Adaptive Integration Approach offers a framework for choosing the former path.
Nick Potkalitsky, Ph.D.
Check out some of our favorite Substacks:
Mike Kentz’s AI EduPathways: Insights from one of our most insightful, creative, and eloquent AI educators in the business!!!
Terry Underwood’s Learning to Read, Reading to Learn: The most penetrating investigation of the intersections between compositional theory, literacy studies, and AI on the internet!!!
Suzi’s When Life Gives You AI: An cutting-edge exploration of the intersection among computer science, neuroscience, and philosophy
Alejandro Piad Morffis’s The Computerist Journal: Unmatched investigations into coding, machine learning, computational theory, and practical AI applications
Michael Woudenberg’s Polymathic Being: Polymathic wisdom brought to you every Sunday morning with your first cup of coffee
Rob Nelson’s AI Log: Incredibly deep and insightful essay about AI’s impact on higher ed, society, and culture.
Michael Spencer’s AI Supremacy: The most comprehensive and current analysis of AI news and trends, featuring numerous intriguing guest posts
Daniel Bashir’s The Gradient Podcast: The top interviews with leading AI experts, researchers, developers, and linguists.
Daniel Nest’s Why Try AI?: The most amazing updates on AI tools and techniques
Riccardo Vocca’s The Intelligent Friend: An intriguing examination of the diverse ways AI is transforming our lives and the world around us.
Jason Gulya’s The AI Edventure: An important exploration of cutting edge innovations in AI-responsive curriculum and pedagogy.
Thanks for pulling together so many different threads related to AI and education. Through the whole read of your piece, I kept nodding yes and thinking about how to bring this to my program and college.
Very helpful and thanks for sharing this AI integration approch for K-12.