The real challenge in AI education isn't whether to use these tools, but how to choose wisely among them. As artificial intelligence becomes embedded in classrooms, simulation centers, and corporate training programs, educators themselves are taking the lead in deciding which tools deserve a place in learning environments. The shift reflects a growing recognition that technology adoption decisions should start with educational purpose, not with what's newest or most popular.

What's Driving the Shift Toward Educator-Led AI Adoption?

For decades, schools and training programs have often adopted educational technology based on vendor promises or institutional enthusiasm. Today, that pattern is reversing. Faculty members, instructional designers, and academic leaders are increasingly responsible for evaluating whether AI tools actually improve learning outcomes. This is especially true in healthcare education, where decisions about technology can ultimately influence patient care and clinical judgment.

Dr. William Monroe, Professor of Professional Practice at Louisiana State University's Paul M. Hebert Law Center, exemplifies this educator-centered approach. Rather than treating AI as something to avoid or to depend on completely, Monroe encourages students to think critically about how these tools influence learning.

"I've been developing an AI literacy thread in my course based upon Anna Mills's Peer & AI Review + Reflection Project (PAIRR). I'm curious about how AI tools influence the work of learning designers and teachers, especially those who are early in their professional lives," Monroe explained.

Dr. William Monroe, Professor of Professional Practice, LSU Paul M. Hebert Law Center

This approach moves beyond simply using AI in teaching practice. Instead, Monroe helps students create meaningful learning experiences grounded in research and evidence. His background spans language education, library science, and educational technology, giving him a unique perspective on how technology reshapes learning.

How Should Educators Evaluate AI Tools for Their Classrooms?

Healthcare educators at institutions like the MGH Institute of Health Professions have developed a practical framework for assessing whether an AI tool belongs in their programs. The framework prioritizes learning outcomes over technological novelty. One of the biggest mistakes institutions make is focusing on technology before educational purpose. An AI tool should support learning outcomes, not distract from them.

Before adopting any AI platform, educators should ask themselves a series of critical questions:

• Learning Impact: Does the tool improve student learning outcomes and strengthen clinical reasoning, communication skills, or other core competencies?
• Real Problem-Solving: Does it solve an actual educational challenge, or does it add unnecessary complexity to existing processes?
• Data Privacy and Security: What data does the platform collect, how is it stored, who can access it, and does the institution maintain ownership of educational content?
• Accuracy and Bias: Is the content evidence-informed, and how easily can faculty identify and correct errors or biased outputs?
• Accessibility and Equity: Does the tool work for diverse learners, including those with disabilities and multilingual students?
• Alignment with Standards: Does the AI system support competency-based education, assessment integrity, and transparent grading practices?

These considerations reflect a fundamental truth: AI systems can produce inaccurate or biased information, and in healthcare education, that risk carries serious implications. Some generative AI systems may provide incorrect answers, incomplete explanations, or fabricated references. Bias can also appear in case examples, assessment recommendations, or predictive analytics.

Steps to Implement Educator-Led AI Evaluation in Your Institution

• Start with Learning Goals: Define what you want students to learn before considering any technology. Ask whether an AI tool supports those specific outcomes or distracts from them.
• Involve Faculty in Decision-Making: Create evaluation committees that include teachers, instructional designers, and subject-matter experts. These professionals understand the real challenges in their classrooms and can assess whether a tool addresses them.
• Establish Clear Data Governance Policies: Before adopting any AI platform, develop institutional policies around acceptable AI use, data retention, student privacy, and how outputs should be verified by human educators.
• Test Tools with Pilot Groups: Rather than institution-wide rollouts, pilot new AI tools with small groups of students and faculty. Gather feedback on whether the tool improves learning and whether it creates unintended barriers.
• Build AI Literacy Among Educators and Students: Help faculty and learners understand both the opportunities and limitations of AI. This critical thinking approach prepares educators to make informed decisions as technology continues to evolve.

What About Privacy and Ambient AI in Classrooms?

A new generation of tools called "ambient AI" is emerging in educational settings. Unlike traditional AI applications that require users to actively enter prompts, ambient AI operates in the background, using data from classroom environments, learning platforms, and student interactions to identify patterns and provide insights. These tools could help teachers understand student engagement, identify struggling students, and personalize instruction.

However, ambient AI also raises significant privacy concerns. Many systems rely on cameras, microphones, sensors, and other forms of data collection to interpret classroom activity. This raises critical questions about what information is collected, who controls it, how long it is retained, and whether vendors can use student data to improve their products. There are also concerns that systems designed to infer engagement, attention, or emotion may produce inaccurate or biased conclusions that could negatively impact students.

The challenges are especially pronounced in the United States, where a decentralized education system means individual states, school districts, and local communities must balance innovation with differing expectations around privacy, transparency, and appropriate technology use. As counties and school systems consider the future of AI integration, ambient AI highlights the need for thoughtful policies that prioritize student outcomes while ensuring responsible data practices, educator involvement, and public confidence.

Why Learning Design Matters More Than Technology

Monroe's approach to multimedia design offers a broader lesson about how educators should think about AI. Many people rely on instinct when creating presentations, videos, or online training. But decades of research on how multimedia affects learning shows that intuitions can be improved by understanding the evidence.

"It's hard for me to imagine a person today who lacks an intuitive sense of how their attention is shaped by the multimedia they interact with daily. And yet, I think that our intuitions can be assisted by looking at how this shaping works to the benefit and detriment of our learners," Monroe noted.

Dr. William Monroe, Professor of Professional Practice, LSU Paul M. Hebert Law Center

This principle applies directly to AI adoption. Effective educational technology, including AI systems, should be strengthened by an understanding of how it influences attention, memory, and learning. By grounding design decisions in evidence instead of assumptions, instructors can create tools that support learning rather than simply capture attention.

The bottom line is clear: as AI tools proliferate in education, the educators themselves must remain in control of the decision-making process. Technology should serve learning goals, not the reverse. Institutions that prioritize educator expertise, evidence-based evaluation, and student outcomes over vendor promises are positioning themselves to harness AI's potential while protecting the human elements that make education meaningful.