From Restriction to Integration: Practical Strategies for Embracing AI in Online Courses – Faculty Focus

A 2025 survey by the Higher Education Policy Institute found that 92% of university students now use AI tools in their studies, up from 66% the previous year (Freeman 2025). For online educators, the question has shifted from whether to allow AI to how to design courses that encourage deep learning, no matter AI’s role. The strategies that follow, based on current research, my classroom experience, and emerging best practices, offer practical ways to incorporate AI into online courses while focusing on real learning. 

Shift the Focus from Output to Process

When we only assess students based on completed essays, reports, or projects, we risk letting AI handle most of the work unnoticed. Kofinas, Tsay, and Pike (2025) highlighted this issue: in a study involving two UK universities, experienced markers struggled to reliably distinguish student work from content generated or altered by ChatGPT, with detection rates as low as 33%.  AI detection tools are equally problematic; Liang et al. (2023) found that seven widely used detectors misclassified over 61% of essays by non-native English speakers as AI-generated. If neither human judgment nor automated detection can reliably identify AI use, we must develop assessment methods that reveal learning progress throughout the process. 

In most of my online courses, I incorporate staged assignments with mandatory checkpoints. Students start by submitting a topic proposal outlining their personal rationale, then progress to an annotated outline with initial sources. Next, they submit a rough draft accompanied by a reflective memo that explains their reasoning and any challenges encountered. The process concludes with a final submission that includes a revision narrative. These checkpoints create a documented record of their intellectual development that cannot be easily fabricated by AI. Additionally, I ask students to video record their final projects. Instead of simply submitting a paper or file, each student records a walkthrough explaining their process, decision-making, and how they addressed specific challenges. Grading this explanation alongside their deliverable provides a clearer view of their true understanding. This approach is especially effective in online courses, where in-person cues indicating comprehension are absent.  

Instead of prohibiting the use of AI, it is more effective to assign tasks that require students to use AI tools and then have them critically assess the outputs. For example, in my programming course, students might use AI to explain existing code or propose solutions to bugs. They are then expected to (a) test each suggestion, (b) articulate in comments why a particular fix works or doesn’t, and (c) reflect briefly on what they have learned about debugging strategies. I evaluate their testing process and reasoning rather than the AI’s initial suggestions. This approach encourages deeper engagement than simply copying and pasting solutions. 

In asynchronous online courses with less frequent instructor interaction, I have adopted a new approach to enhance engagement in weekly discussions. I ask students to use AI tools to generate practice questions and sample answers, allowing them to self-assess their understanding. Students then post their AI-generated questions and answers as original discussion posts, reflecting on which questions were most helpful and identifying any gaps in the tool’s knowledge. Additionally, they evaluate at least two other question/answer sets created by their peers. This method fosters a peer dialogue focused on critical assessment, reducing the instructor’s workload in creating every quiz while encouraging collaborative learning. 

Enhance Engagement via Collaborative Learning

Online courses often lead to student isolation, and the use of AI can worsen this if students rely solely on chatbots rather than engaging with peers. Research indicates that AI tools are most effective when used as part of interactive teaching strategies such as project-based learning and scaffolded feedback (Long et al. 2026). The technology thrives when it complements human interaction rather than substitutes it. 

Collaborative learning, where students work together toward shared academic goals, is among the most effective strategies for enhancing online engagement (Johnson, Johnson, and Smith 2007). In my courses, I incorporate various collaborative activities that naturally discourage the use of AI shortcuts. For instance, jigsaw activities assign each student a specific component of a topic to master and subsequently teach to their peers; the group then synthesizes this information into a cohesive whole. Another example is collaborative peer programming projects, which require teams to divide roles, establish working agreements, and develop the project collectively, with accountability embedded in each role. Finally, structured peer review assignments further promote engagement by prompting students to evaluate a classmate’s work and provide constructive feedback before final submission. 

Integrate AI Literacy into the Curriculum

If students are to utilize AI effectively, it is our responsibility to teach them responsible usage. Incorporating AI awareness modules into our classes can equip students with clear guidelines for appropriate and inappropriate uses, such as seeking hints versus copying complete solutions. This framing can be adapted to the introductory modules of any online course. Each semester, I dedicate one early class session to what I call “AI transparency training.” In this session, students experiment with Generative AI tools such as ChatGPT on low-stakes assignments, then reflect on the AI’s accuracy, its errors, and their own contributions that the AI could not replicate. This single exercise helps transition the classroom environment from secrecy to transparency and sets expectations for the rest of the semester. 

Focus on Making “Learning” the Main Goal

The core principle behind these strategies is clear: prioritize evaluating the learning process over just the final product. While AI has made producing polished outputs easier, it hasn’t replaced the ability to assess reasoning, collaboration, and explanation. As Kofinas, Tsay, and Pike (2025) noted, generative AI has increased the importance of social learning, since explicit knowledge can now be quickly reproduced by machines. What remains uniquely human, and therefore fundamentally valuable to assess, is how students apply, question, and communicate their understanding. 

The online classroom doesn’t need to be a setting where AI diminishes learning. Through intentional design, it can be transformed into a space where AI supports and improves the learning process. Although this transition requires effort, it results in a learning experience that is more challenging and better aligned with the real-world students are preparing to enter.

Taoufik Ennoure, PhD, is a Computer Science Educator and Researcher with over 19 years of experience in Higher Education. He is an Associate Professor at the Community College of Philadelphia and an Adjunct Professor of Computer Science at NYIT, Monroe University, and Baruch College, City University of New York. His research interests include Algorithm Optimization, Artificial Intelligence, Big Data Analytics, and online learning pedagogy.