Key points:
- To use AI well in instruction, educators may need to loosen perfectionist habits
- Aligning AI with pedagogy, privacy, and outcomes
- AI as a scaffold for learning: From access to judgment
- For more news on AI and learning, visit eCN’s AI in Education hub
In my Systems Analysis and Design course, students are not handed the requirements for building a software application. They have to uncover them by asking the right questions within an AI-based learning activity. Because identifying user requirements is a critical analytical skill, the assignment turns that process into the lesson itself. The goal is not to teach prompting for its own sake, but to use AI to support an instructional task that still demands precision, judgment, and accuracy.
The activity mirrors a live requirement-gathering session. Students must ask precise questions to get useful answers. Broad questions produce broad responses, whether the source is AI or a person. In that sense, AI does not create ambiguity. It exposes it. When students ask weak or incomplete questions, the problem often lies in their thinking rather than in the tool itself. The exercise also shows how so-called hallucinations can sometimes reflect user error rather than AI failure.
In the past, I would have taught this through a live classroom exercise. Now, AI can mediate the questioning and analysis while still supporting the learning objective. Technology-mediated instruction must pass an equivalency test: The artifact may differ from what happens in a classroom, but the outcome must remain the same. It is equivalent rather than equal because the process changes while the result does not. AI helped me imagine and shape the design, but I remained in control throughout. The result was a learning experience in which each student’s participation was essential and cheating became impossible because the real work was not simply finding answers, but learning how to think, ask, synthesize, and refine. A year ago, I would not have thought this was possible.
That classroom example reflects a broader shift in education. As natural language tools lower the technical barrier, the central challenge is no longer mastering software or technical routines, but exercising instructional judgment: deciding what to build, how to structure learning, and how to use AI as a partner in design. Creativity requires imagination, but it also requires ideas with practical value. The deeper institutional need, then, is not simply tool adoption, but the capacity for creative instructional design. AI platforms differ in scope and complexity, and faculty need guidance, training, and informed leadership. The pace of change is too rapid for individual instructors to navigate alone. Institutions must help translate technological change into practical, pedagogically sound choices.
To build that capacity, institutions need leaders who actively use the tools they recommend and bring more than surface-level familiarity. There is no substitute for firsthand experience when anticipating faculty needs in this emerging mode of instruction. Many institutions have appointed a point person, but what makes that role effective is not the title. It is sustained research, practical expertise, and ongoing engagement. Simply directing faculty to websites, videos, or other resources is markedly insufficient. Because AI information changes so quickly, faculty need ongoing conversation, interpretation, and guidance from someone who can help them make sense of it. The goal is not just tool adoption, but stronger institutional support for creative instructional design.
For instructors, the central challenge has shifted from the technical to the creative. Like artists facing a blank canvas, they must decide what to create from an open-ended starting point, and that can feel unsettling. One of AI’s greatest strengths is its ability to support planning through conversation. Many faculty do not yet realize that some tools can help them think through a design before anything is built. Learning to use that capacity well is essential. It can reduce wasted tokens, bring structure to a project before development begins, and prevent unnecessary rework later. Approaching AI as though speaking with an expert colleague can generate direction, structure, and new ideas. Over time, instructors will learn to plan more effectively, ask better questions, and use AI to support creative work. That kind of thinking is often more demanding than simply following a checklist.
The challenge AI presents for many faculty is not simply learning a tool, but adapting to a more open-ended way of thinking. Faculty are trained to value precision and correctness, which can make the open-ended nature of AI feel uncomfortable. Yet learning to work with these tools is not rooted in technical proficiency. It is about using AI as a partner in exploration and learning to ask the kinds of “what if” questions that often lead to stronger ideas. To use AI well in instruction, faculty may need to loosen perfectionist habits, accept some trial and error, and recognize that the most important tool is still their own creative judgment.
If instructional practice now demands more creative thinking, institutions must plan for that shift. Individual instructors may eventually create their own materials, but in the early stages, shared content will likely be necessary because this approach is new to everyone. Asking faculty to learn new tools, processes, and methods while also maintaining effective instruction is simply too much. With only 168 hours in a week, time must be treated as a limited and valuable resource. Building institutional capacity for creative instructional design requires time, support, and shared development structures that make this work possible.
Choosing a single AI tool or platform is also not easy. Larger providers such as Google, Microsoft, OpenAI, and Anthropic offer broad and capable ecosystems, while smaller platforms such as Base44 may offer greater simplicity. Starting with one ecosystem is a practical approach because it allows faculty to focus on core concepts that can later be transferred to other tools and platforms. It is better to begin with a manageable scope than to take on too much at once, which is a common mistake when people try to evaluate many tools simultaneously.
Cost remains another major challenge in AI adoption, affecting both instructor development and student access to learning materials and targeted AI use. Although AI models are becoming more affordable, the expense is still substantial. In the current funding environment, this poses a serious obstacle for many institutions. In higher education, passing those costs on to students is also problematic. Equitable access must be built into planning from the start rather than left unresolved. AI may be able to generate Santa Claus, but it still cannot provide free tokens or credits.
Beyond the logistical challenges of AI, the deeper shift is toward using it as a cognitive partner in creative work. Most instructors will need to devote more time to planning, instructional design, and formative feedback. AI can quickly generate content and presentation layers, but the harder work lies in shaping the ideas, structure, and learning experience those outputs are meant to support. Creating materials may become easier, but deciding what to create and refining it through iteration will require more thought.
Even so, one of the primary concerns in this shift remains cheating. That issue must be addressed in the design of assignments and instructional materials. Asking AI for ideas on how to reduce opportunities for cheating may seem circular, but it can be effective. A back-and-forth dialogue with AI can help instructors develop design choices that make dishonest shortcuts less appealing or less useful.
At the same time, focusing too heavily on preventing cheating can drain time and attention. We are entering an era in which answers are readily available, including those that involve reasoning, not just recall. Instructors cannot succeed by becoming integrity police, so a different emphasis is needed. The better focus is learner agency. We cannot force students to learn; we can only encourage and guide them. Students who want to learn will have access to resources, feedback, and support on an unprecedented scale. Those who do not will likely turn to some form of cheating. In higher education, those gaps in learning eventually surface through assessments, certification exams, licensing exams, and other measures of achievement. Responsibility will be shared, and students will have real stakes in the outcome.
In K-12 education, the challenge is especially significant because learning experiences must still build deep understanding of essential concepts without overreliance on AI. Memorizing multiplication tables, for example, remains important for higher-level math, yet students have long used calculators and education has adapted. Distinguishing between cheating and efficiency will require ongoing research and careful judgment. The boundary between the two is becoming harder to define.
We are learning in real time. The future of teaching will demand more than content delivery or technical fluency. It will require instructional judgment in the design of learning. That challenge is significant, and it is also institutional. Schools must build the capacity to support creative instructional design, not simply adopt new tools. The pace of change is accelerating and will reward those who adapt, experiment, and lead. The challenge is no longer whether to adapt, but how well we will do it.
Dr. Mark Taormino is a professor in the Computing & Information Technology Department at the College of Southern Nevada. His research interests are focused on Instructional Technology and using technology as an instructional and learning catalyst. In addition to teaching computer and technology courses, he works closely with faculty to integrate pedagogy and technology into teaching practice.
- The real work of AI and instructional technology is creative - June 26, 2026
- How smart campuses are turning efficiency into growth - June 24, 2026
- For schools, cyber resilience starts at the data layer - June 22, 2026