AI in Education: Research, Practice, and Perspectives Mini Conference

This qualitative practitioner research study explored the attitudes and perceptions of administrators and teachers in a high-performing Midwestern school district regarding the integration of artificial intelligence (AI) in K-12 education.

While AI has rapidly emerged as a transformative force in society, its implementation in schools remains inconsistent, with limited policies, training, and usage.

Semi-structured interviews with nine educators (three principals and six experienced teachers) revealed widespread interest in AI, but also significant uncertainty and uneven adoption.

Educators reported using AI for administrative and instructional tasks, such as grading and resource generation, yet expressed concern over student overreliance and diminished critical thinking.

Findings highlight a lack of formal policies, minimal familiarity with AI tools, and sporadic professional development opportunities.

In response, the researcher, along with colleagues, initiated immediate changes to university-level course content and began offering both University faculty and K-12 educators professional development on AI.

This study emphasizes the urgent need for comprehensive educator preparation in AI integration both at the preservice and professional levels and suggests future research expand to more diverse contexts to inform broader policy and practice.

Presentation File

Increasingly, K-16 teachers and teacher educators face pressure to adopt and adapt new technologies, from learning management systems to digital assessment tools to assistive technologies and other artificial intelligence applications (Cone, 2023). Whether this pressure comes from administrators, peers, the companies themselves, or a fear of being left behind, it can obscure the reality that these platforms, technologies, and their associated literacies are not neutral but rather serve to reinforce or disrupt existing hierarchies (Chander & Krishnamurthy, 2018).

This chapter explores digital platforms in teacher education using a critical media epistemology (CME) lens (Lathrop, 2025). CME is a theoretical framework that integrates aspects of critical media literacy (Kellner & Share, 2019) with epistemic cognition (Sandoval et al., 2016), a field of study focused on how people think about knowledge.

The chapter identifies epistemic issues related to platform use in teacher education and shows how the application of CME could provide a framework for teacher educators to help preservice teachers think more critically, skeptically, and epistemically about the digital platforms they are asked to adopt.

Presentation File

As generative artificial intelligence (GenAI) becomes increasingly embedded in instructional planning, teacher education programs must move beyond efficiency-oriented adoption toward equity-centered, pedagogically grounded integration. This mixed-methods longitudinal study examines a year-long, scaffolded approach to integrating GenAI within an early childhood teacher residency program, with a specific focus on culturally responsive teaching (CRT) instructional planning. Guided by the AI/I-TPACK and CRT frameworks, fourteen bilingual teacher candidates (TCs) participated in this study. Data sources included pre- and post-surveys of CRT self-efficacy, lesson plans and instructional artifacts, reflective coursework, and focus group interviews. Quantitative analyses revealed statistically significant gains in TCs’ CRT self-efficacy across all CRTL standards, particularly in areas related to instructional enactment, differentiation, and student-centered planning. Qualitative findings indicate that while GenAI tools supported efficiency, multilingual scaffolding, and differentiation, they frequently lacked cultural specificity and reproduced dominant representations, requiring TCs to engage in critical evaluation, revision, and contextualization. Findings suggest that scaffolded, critical engagement with GenAI and CRT, within this context, coincided with stronger culturally responsive instructional planning. The study contributes a replicable design model for equity-oriented GenAI integration in teacher education and offers implications for preparing early childhood educators to use GenAI as a reflective pedagogical tool rather than a substitute for professional judgment

Presentation File

As generative AI tools become embedded in writing instruction, educators face urgent questions about how these technologies encode and reinforce language standardization — particularly for multilingual and multidialectal students.This practitioner reflection describes a series of AI-powered "remix" activities developed and refined in corequisite writing labs.

Building on a prior action research study, this paper turns to the pedagogical questions that emerged after the study: how AI tools might be used not as writing shortcuts but as objects of critical inquiry.

Drawing on frameworks including critical literacy, plurilingualism, culturally relevant pedagogy, and AI literacy, we describe four remix activity types — text to image, genre transformation, tone adjustment, and translation comparison — each anchored by a critical question about AI's assumptions and values.

Across these activities, students examined how AI systems privilege certain language varieties, surfaced tacit knowledge about genre and audience, and engaged in substantive discussions about language power and standardization. We argue that AI tools, approached critically, can become productive sites for developing the metalinguistic awareness students need to make informed rhetorical choices rather than simply submit to the judgments encoded in the tools they use
Presentation File

As National Louis University seeks to plan and advance its way amid the potential perils and pitfalls of so-named artificial intelligence, one finds that with reasonable workflows and guard rails, the use of AI proposes both potential promise and prospect, to assist and augment the natural intelligence that students and teachers bring to the classroom, encouraging them to pursue STEM-related activities with the preparation and deployment of focused and directed AI-bots.

This paper presents examples and a case study of that directed and focused approach, the incremental lessons learned to date, and a proposed path to advance worthwhile AI usage.
Presentation File

The rapid adoption of generative artificial intelligence in K-12 schools has outpaced educators' capacity to evaluate its effects on learning. This conceptual article argues that the dominant institutional responses treating AI as either an integrity threat or a productivity tool share an unexamined assumption that the fundamental nature of learning remains unchanged.

Drawing on Walter Ong's media ecology and extending it through the concept of tertiary algorithmicity, the article offers a theoretical framework for understanding generative AI as a qualitative shift in the symbolic environment of schooling. For the first time, the prevailing communication technology ctively generates symbolic content, enabling what Kapur's productive failure research identifies as unproductive success: competent performance without underlying cognitive development rather than just mediating symbolic content.

The article introduces the Pedagogical Friction Framework, a four-dimensional model (noetic, rhetorical, existential, and infrastructural) for designing learning environments that preserve the productive cognitive struggle essential to genuine understanding. Implications for curriculum and pedagogy, institutional leadership and policy, and the student experience of learning are examined from a practitioner perspective grounded in K–12 technology leadership.

Presentation File

This autoethnographic study examines how generative AI has entered higher education, reshaping teachers’ and students’ inner lives, relational practices, and institutional ecologies. Drawing on the author’s embodied experiences within the higher education landscape, the analysis consists of plugging in theory, research, policy, and current events as a means to connect individual experience with broader concepts, events, and structures (Jackson & Mazzei, 2022).

The paper maps affective responses such as awe and anxiety, pedagogical tensions such as assessment, transparency, and trust, and political-economic drivers such as capitalism and policy. The study foregrounds methodological reflexivity, arguing that critical autoethnography can surface situated knowledges and prompt timely ethical action.

The paper calls for educator-led, context-sensitive research and policy interventions to preserve humane pedagogies as AI becomes entwined with learning.
Presentation File

This article examines how faculty at a college of education are integrating generative AI tools into graduate and doctoral educator preparation programs. Drawing on faculty inquiry and course-based observations across multiple contexts — including early childhood education, secondary education, special education, and doctoral programs in education — the article synthesizes emerging lessons and practices of AI literacy development, pedagogical approaches, and instructional design.

We present an evidence-informed, reflective account of faculty practice through the lens of the scholarship of teaching, exploring how educators are navigating AI content, literacy, and practices in real courses, what students are learning, and what instructional decisions support thoughtful, critical AI use. Three shared themes emerge: the importance of developing professional judgment alongside technical familiarity with AI, the necessity of addressing equity and ethics explicitly rather than incidentally, and the value of positioning AI as a thinking partner and collaborator rather than a replacement for human expertise.

The article concludes with practical implications for faculty seeking to integrate AI in ways that are forthright about uncertainty, grounded in pedagogical purpose, and attentive to learners' diverse needs.
Presentation File

In an era of instant AI answers, how do we ensure students are still doing the "heavy lifting" of learning? This session explores a shift in the classroom dynamic: moving students from passive recipients of AI output to active "Prompt Engineers" and "Learning Architects."

Drawing on the principles of Project-Based Learning (PBL) and the GRASP model (Goal, Role, Audience, Situation, Product), we will discuss how to leverage AI for lower-order tasks—like initial research and drafting—to create more space for high-order critical thinking.

Participants will discover how to help students use AI to iterate, self-reflect, and defend their work, turning a potential shortcut into a powerful tool for mastery and metacognition.
Presentation File

Artificial intelligence is transforming education, but most of the discussion has centered on what AI can do. This presentation asks a different question: what does AI do to educational time? Generative AI is not just a new pedagogical tool but a temporal force, one that compresses the time horizons within which students work, think, and form habits of mind. This paper introduces the concept of temporal ecology: the patterned conditions under which action unfolds in educational environments, including the degree of order or fragmentation, the protection of duration, and the intelligibility of effort across time.

When used as a substitute for student work, generative AI produces output that is immediate, frictionless, and polished. This is opposed to the slower processes through which genuine learning occurs. Used uncritically, AI functions as a high-time-preference technology, privileging the shortest route from prompt to product and displacing the developmental work that happens between a question and a hard-won answer. The problem is not only academic dishonesty. It is the erosion of conditions that make patience, revision, and sustained effort meaningful.

Drawing on behavioral economics and developmental psychology, the presentation argues that future-oriented action depends on the temporal structure of the environment itself. Montessori education and apprenticeship models serve as illustrative counter-models, clarifying what is at stake in the temporal organization of our educational institutions.

Participants will be able to identify generative AI as a temporal force and explain how its structural features potentially erode the educational conditions that make patience, revision, and sustained effort meaningful.

Participants will gain a working understanding of temporal ecology as a framework for evaluating their own classrooms and institutions, using Montessori education and apprenticeship models as reference points for protecting educational time in an increasingly AI-connected landscape.
Presentation File

After more than half a century in education, AI has presented new opportunities and challenges. How can we assist our students to use it ethically and responsibly without dissuading them from using it to benefit learning in classrooms of all types.

The intent is to suggest opportunities to ethically integrate AI in teaching and learning
Presentation File

AI tools are changing how students do their work. Today, students can use AI to generate essays, solve problems, and organize ideas in seconds. Because of this, the final product—an essay, a paragraph, or an answer—no longer clearly shows what a student actually understands. This makes it harder for teachers to know what students are learning and where they need support. This session focuses on a key question: If students can use AI to produce the final answer, what should we be assessing instead—and why?

I argue that we need to shift our focus from the final product to the process of learning. This means paying attention to how students think: how they plan their work, how they make decisions, how they use AI, and how they reflect on their understanding. These parts of the process matter because they show whether a student is truly learning or simply relying on a tool.

Research on metacognition and self-regulated learning supports this shift. When students are asked to explain their thinking, reflect on their choices, and monitor their own learning, they develop deeper understanding and become more independent learners. At the same time, research on AI use shows that without clear structure, students may skip important thinking steps and rely too heavily on AI, which can limit their learning.

In this session, I will share a few practical ways to redesign assessments so that student thinking becomes more visible. For example, asking students to show their planning, explain how they used AI, and reflect on what they learned. These changes are simple, but they help teachers better understand student learning and help students stay engaged in the thinking process. The goal of this session is not to avoid AI, but to use it more intentionally. By focusing on process, decision-making, and reflection, we can design assessments that better reflect real learning in a world where answers are easy to generate.
Presentation File

Central Question:
How can agentic AI be leveraged to help teacher candidates master the logic and content of multiple-choice licensure exams across early childhood, elementary, and secondary education?

Overview & Focus:
High-stakes multiple-choice exams often measure more than just rote memorization; they require candidates to apply pedagogical theories to "best-practice" scenarios where multiple answers may seem plausible. This practice-focused session introduces an AI Licensure Co-Pilot—an agentic tutor designed to move beyond simple flashcards. Rooted in Cognitive Load Theory, this agent helps candidates deconstruct complex multiple-choice stems, identify "distractor" patterns, and strengthen the specific content knowledge required for Early Childhood, Elementary and Secondary licensure.

What will be covered:
The session will demonstrate how AI agents can provide personalized, real-time feedback on practice questions and content.. We will discuss the framework for grounding AI in specific state standards (e.g., ILTS) to ensure technical accuracy and the pedagogical shift from "giving the answer" to "teaching the rationale."

Key Takeaways:
Participants will learn how to transform static performance data—including pre-tests, diagnostics, and official exam reports—into a dynamic Adaptive Study Roadmap. By leveraging AI to identify patterns in missed questions, students leave not just with a list of "what" to study, but a prioritized, time-bound "how" that evolves with their progress.

A roadmap for using AI to create infinite, standard-aligned multiple-choice practice scenarios. Techniques for using AI to analyze a candidate’s "error patterns" across different developmental domains. Strategies to help candidates overcome the linguistic and cognitive hurdles of standardized testing.

Briefly describe the AI tools, interactive components, or activities participants will experience.

AI Tools:
Participants will engage with a custom Education Test-Prep Agent (built on GPT-4o). This agent is specifically tuned to analyze multiple-choice questions, provide Socratic hints rather than immediate answers, and explain the "why" behind correct and incorrect options based on K-12 pedagogical standards.

Interactive Components:
The "Distractor" Deconstruction: Participants will provide a sample multiple-choice question to the AI. Together, we will prompt the agent to explain why a specific "distractor" answer is incorrect according to developmental theory (e.g., "Why is this intervention inappropriate for a 2nd-grade Tier 1 setting?").

Reverse-Engineering Lab:
A hands-on activity where attendees ask the AI to generate three different versions of a question based on a single learning objective (Early Childhood vs. Elementary) to see how the exam's complexity scales.

The "Hint-First" Simulation:
A live demo where the AI acts as a tutor that refuses to give the letter answer (A, B, C, D) but instead provides three "clues" rooted in teaching standards to help the participant arrive at the answer themselves.

Ethics & Accuracy Audit:
A quick group evaluation of an AI-generated explanation to identify any potential hallucinations or misinterpretations of state-specific teaching laws.

The "Concept Clarity" Pivot:
If a participant is unfamiliar with a specific term or theory within a question, the AI can pivot from "testing mode" to "teaching mode," providing a concise, 1-minute breakdown of the core content before returning to the practice exercise.

Many educators use AI for basic tasks like proofreading or drafting emails, but its capabilities extend much further. This session helps educators move from surface-level AI use to building practical, interconnected workflows that truly save time. We’ll focus on Google's AI tools including Gemini features like Canvas, Gems, Deep Research, and NotebookLM integration and how to connect these tools seamlessly. We will walk through real-world instructional scenarios from start to finish, showing how to truly partner with AI in a workflow that gets you great results and improves teaching and learning meaningfully.

Interactive Components

This session centers on a live, step-by-step demonstration of a complete instructional workflow. Participants will follow along as we tackle a real teaching scenario using Google's AI ecosystem. We will use Gemini's research mode to gather topical data, collaboratively build custom Gems tailored to specific grading rubrics, utilize NotebookLM to synthesize sample course literature, and leverage Gemini's Canvas to draft and edit lesson materials. Attendees are encouraged to bring their own upcoming course topics to practice applying these workflows in real-time.

Presentation File

In North Shore School District 112 in north suburban Chicago, one school district's approach to incorporating generative artificial intelligence went from urgency (the need) to agency (the approach). Learn how hands-on usage of AI by students (grades 4, 6, and 8) led to approved policy and guidance in a relatively short period of time. Our journey and steps can be replicated and modified to any setting. We'll share illustrations of Playlab AI, Notebook LM, Magic School AI, and our new policy and our approaches to AI Literacy, Fluency and Application in our PK-8 school district.

AI tool

Playlab AI, Notebook LM, Claude AI, Magic School AI

Presentation File

AI adoption in education has outpaced the intentionality educators need to use it well. The result is a set of predictable, understandable mistakes — not born of carelessness, but of pressure, novelty, and thin professional support. Some educators hand off too much and accept whatever comes back. Others find one tool, one use, and stop there. Others cycle through every new option without landing anywhere useful. What these patterns share is the absence of a clear evaluative framework.

The central question: When does AI add value to your practice — and how would you know if it did?

The framework is a practitioner-developed "Top 10 Mistakes" list, organized around a core claim: without intentionality and sufficient background knowledge, educators cannot prompt effectively, evaluate output critically, or determine whether AI is improving anything at all. Fluent-sounding output is not the same as useful output.

Both presenters work from their own institutional contexts — one in K–12, one in higher education — and the contrast between them is built into the session as a feature, not just a framing note.

The workshop draws on experiences from two contexts anchored to specific mistakes from the list, illustrating how concrete shifts in purpose and process change what AI can actually contribute. Participants leave with a living document combining the proposed top 10 mistakes with group additions and a clearer sense of where AI earns its place in their own practice — and where it doesn't.

Briefly describe the AI tools, interactive components, or activities participants will experience.

Participants will engage with AI both as the subject of analysis and as a live tool, moving between reflection and hands-on experimentation. Interactive components include:

• Opening poll: Participants respond to 3–4 items from the Top 10 list before it's revealed, generating live data and establishing a "we've all been here" tone from the start

• Chat blast and forced-choice reactions woven into the Top 10 countdown — quick, low-stakes responses that surface patterns and disagreements across institutional contexts

• Live prompting activity: Participants begin with a deliberately vague prompt — "make a lesson plan for my course" — run it through any AI tool (or multiple), then refine using their own expertise: course level, learning objectives, student context. The delta between the two outputs makes the argument concrete. Structured with time to think first, then as breakout rooms (split by K–12 and higher ed) if the group size allows, or as a large-group demo with live audience participation if not. Discussion afterwards centers on the need for thoughtful prompts, evaluating, and refining as well as different benefits of different tools

• Cross-sector exchange: a brief structured conversation between presenters on where K–12 and higher ed diverge, with audience weighing in via chat

• Collective list-building: participants contribute mistakes they've observed in their own contexts, extending the Top 10 into a living document generated by the room

• Closing reflection: each participant names one mistake that resonates and one shift they'll consider — dropped in chat and briefly discussed before close

Presentation File

This workshop focuses on elevating instructional rigor through purposeful AI integration, grounded in Bloom’s Taxonomy and high-quality task design. The central question guiding this session is:

How can educators design AI-supported learning experiences that deepen student thinking rather than replace it?

Participants will explore the distinction between low-rigor AI use, where technology substitutes for student thinking, and high-rigor AI use, where AI is leveraged to generate cognitive challenge. Using a practical framework that aligns AI use with higher-order thinking skills, the session will model how to shift from tasks such as simple summarization or answer generation to more complex practices like argumentation, error analysis, and evidence-based reasoning.

The workshop is rooted in both research and practice, drawing on principles of cognitive rigor, revised Bloom’s Taxonomy, and emerging best practices in ethical and effective AI integration in education. It emphasizes instructional decision-making, task design, and equity, ensuring that AI enhances, not diminishes, student learning opportunities.

Briefly describe the AI tools, interactive components, or activities participants will experience.

Participants will engage in a highly interactive, hands-on experience using generative AI tools (such as ChatGPT, Gemini or Claude) to design and refine instructional tasks.

Key activities include:

• Prompt Rewriting for Rigor: Participants will transform low-level AI prompts into high-rigor prompts that require analysis, evaluation, and creation.

• AI Prompt Bank Development: Educators will create their own bank of prompts aligned to higher-order thinking, including:

• Generating evaluation-level questions

• Designing authentic performance tasks

• Creating flawed arguments for critique

• Developing error analysis activities

• Building structured academic debates

• Task Design Challenge: Participants will apply the framework to redesign an existing lesson or assignment, integrating AI in ways that promote critical thinking and student ownership.

• By the end of the session, participants will leave with practical tools, ready-to-use prompts, and a clear framework for integrating AI in ways that meaningfully elevate rigor and student engagement.

Presentation File

This session presents an innovative project that explores the development and implementation of a custom Large Language Model (LLM) designed for maternal-child health education. The presentation will discuss how the presenters leveraged an LLM platform to create a domain-specific AI chatbot trained on program materials and academic resources to deliver a scholar-practitioner experience for students.

The session will highlight how this tailored LLM is embedded directly into courses to provide equitable and personalized learning support. Participants will learn how the chatbot can guide students through assignments, scaffold critical thinking, and offer consistent, just-in-time academic assistance to promote persistence and engagement. Beyond program-specific benefits, the project is structured to assess broader outcomes, including improvements in academic performance, inclusivity, and accessibility in higher education.

The presenters will share and answer question on how to use the approach used to develop the material education LLM for scalability and the development of an implementation framework and guide that can be adapted across disciplines and programs, positioning the approach as a model for institutional-level adoption. Attendees will gain insights into how custom LLMs can advance equity, enhance student learning, and expand access to AI-enabled academic support within diverse educational settings.

Briefly describe the AI tools, interactive components, or activities participants will experience.

Participants will have an opportunity to see how LLMs can be used to instruct and assess student knowledge. Participants will be provided links to use several LLMs.

Presentation File