{"id":112,"date":"2026-07-30T19:32:40","date_gmt":"2026-07-30T19:32:40","guid":{"rendered":"https:\/\/pressbooks.ulib.csuohio.edu\/usingaiinacademics\/?post_type=part&p=112"},"modified":"2026-09-21T18:33:50","modified_gmt":"2026-09-21T18:33:50","slug":"5-ai-in-student-learning","status":"publish","type":"part","link":"https:\/\/pressbooks.ulib.csuohio.edu\/usingaiinacademics\/part\/5-ai-in-student-learning\/","title":{"rendered":"5. AI in Student Learning: Pedagogical Integration, Student Agency, and Institutional Responsibility"},"content":{"raw":"Navid Goudarzi, Dakota King-White, Fred Smith, Alex Sukhoy, Adam Voight\r\n

Abstract<\/h2>\r\nArtificial intelligence (AI) has become an embedded feature of contemporary academic life, shaping how students research, write, study, and prepare for professional environments. This paper articulates an instructional framework for integrating AI into higher education that prioritizes learning outcomes, ethical reasoning, and human judgment. Drawing on Cleveland State University instructional practice, the framework distinguishes between instrumental and executive uses of AI and argues for faculty-guided structures that support transparency, critical evaluation, and workforce readiness.\r\n

1. Introduction on Research on how AI is Impacting Student Learning<\/h2>\r\nGenerative AI has rapidly become embedded in how college students research, write, solve problems, obtain feedback, and prepare for professional work. Because students will encounter these tools both inside and outside the university, treating AI primarily as a new form of academic misconduct is no longer an adequate institutional response. The more consequential question is what students must learn to use AI productively, ethically, and critically. Higher education must prepare students not merely to generate competent outputs, but to exercise judgment about when AI is appropriate, evaluate the accuracy and limitations of its responses, recognize bias, protect sensitive information, and remain accountable for the work they produce. These capabilities will increasingly shape graduates\u2019 readiness to compete in the workplace and contribute responsibly to their professions and communities.\r\n\r\nEmerging research suggests that AI\u2019s effects on learning depend substantially on how it is used. Hon\u2019s (2026) systematic review of 41 studies found that generative AI can increase engagement, provide personalized feedback and explanations, improve efficiency, and support academic performance. At the same time, the review identified risks of cognitive offloading, superficial engagement, overreliance, and diminished critical thinking when students accept AI-generated content without scrutiny. Pallant et al.\u2019s (2026) analysis of 192 student reflections similarly found that students who used AI to construct and augment their own knowledge demonstrated higher grades, applied knowledge, critical thinking, and learning autonomy. Students who used it procedurally or simply reproduced its outputs demonstrated poorer outcomes. Although this evidence remains developing, it indicates that AI is neither inherently beneficial nor inherently harmful to learning. Its educational value depends on whether students possess the skills and dispositions needed to use it as a resource for thinking rather than as a substitute for thinking.\r\n

2. Instrumental vs. Executive AI Use<\/h2>\r\nInstrumental AI use supports learning, while executive AI use replaces cognitive labor. Effective pedagogy bridges this gap by centering evaluation and judgment with the student.\r\n\r\nUse of generative AI tools by students may have both positive and negative effects on student learning. Use of AI to automate routine or lower-level tasks (e.g., checking grammar; doing calculations) has the potential to free time and cognitive resources to engage in higher-level tasks, increasing learning. However, if AI is used for those higher-level tasks, the potential benefit for learning may be lost.\r\n\r\nIn this section, we distinguish two ways in which generative AI tools might be used by students; describe the cognitive consequences of those uses; and provide some examples of the types of research conducted to reach these conclusions.\r\n

Ways Students May Use Generative AI Tools<\/h3>\r\nVarious researchers have distinguished between use of generative AI to support learning and use of AI to replace learning (e.g., Slade, 2025), and there is evidence that students use AI in each of these ways. Student learning may be supported by using AI tutors, obtaining feedback from AI systems on the quality and completeness of an essay, or using AI to corroborate the solution to a problem (see, e.g., Jose et al., 2025). These uses have been characterized as instrumental (Aguilar et al., 2025); Contractor and Reyes (2026) labeled students who use AI this way as augmentation users. When students engage in such uses of AI, they engage in the cognitive processes that support learning.\r\n\r\nWhen students use AI to write an essay or solve problems with little or no cognitive engagement, the opportunity for learning is reduced. Aguilar et al. characterized such uses of AI as executive uses; Bauer et al. (2025) call this an inversion effect; Contractor and Reyes (2026) labeled students who use AI this way as automation users. Surveys show that students report engaging in both types of AI use (e.g., Aguilar et al.), as do content analyses of logs of student interactions with generative AI tools during experimental studies in which use of AI tools is permitted (e.g., Contractor & Reyes, 2026). Because durable learning depends on students generating explanations, retrieving knowledge, organizing information, and solving problems themselves, replacing these processes with AI reduces opportunities for learning.\r\n\r\nInstrumental use of AI represents selective cognitive offloading: routine operations are delegated while students retain responsibility for the cognitive processes central to learning. Executive use of AI involves offloading those very processes responsible for constructing knowledge.\r\n

Consequences of AI Use<\/h3>\r\nInstrumental use of AI has the potential to promote learning because while using AI tools in these ways, students should be engaged in the cognitive processes that underlie learning while using AI to support their work. In contrast, because executive use essentially outsources those aspects of the task that engage those cognitive processes, opportunities for learning are reduced.\r\n\r\nIn numerous studies, students have been asked to report on their use (and misuse) of AI, their engagement with their academic tasks, and their perceptions of the effects of AI on their learning (e.g., Aguilar et al., 2025; Chirkov et al., 2026; Farhat, 2026). Although \u201cthere is no robust body of work on actual learning\u201d (Cooper, 2026, p. 37), a body of work is developing in which researchers are investigating effects on learning of using various AI tools. Although many of these studies are methodologically weak (see, e.g., Bauer et al., 2025), the several studies described here, although imperfect, exemplify rigorous experimental investigations of the effects on learning of AI use and illustrate variability in findings concerning learning.\r\n\r\nBastani et al. (2025) concluded that access to generative AI during practice of math problems negatively affected learning. They randomly assigned math classrooms with close to 1000 Turkish high school students to three experimental conditions that differed according to a resource that was available while students practiced a math concept for 90 minutes. The conditions were access to ChatGPT, access to a specialized ChatGPT tutor that would provide hints and guidance and permit checking of answers, and a no-AI control. The tutor condition constrained AI use (encouraging instrumental AI use), whereas unrestricted access to ChatGPT permitted executive AI use. Following the practice session, students took a closed-book exam. Although students in the ChatGPT conditions performed better on the practice problems than those in the no-AI control condition, these benefits did not carry over to performance on the exam. On the exam, performance of students in the no-AI control condition exceeded that of those in the ChatGPT condition and did not differ significantly from those in the ChatGPT tutor condition. Bassner et al. (2026) conducted a similar study using a programming task with undergraduate- and master\u2019s-level management students, and reached similar conclusions: Although students in the AI conditions outperformed those in the no-AI control condition on the task, this advantage was not reflected in better performance on the measures used to assess learning. This replicated an important pattern observed by Bastani et al.: AI improved immediate performance without corresponding improvements in learning.\r\n\r\nContractor and Reyes (2026) tasked students at an elite college with learning about and writing an essay on an unfamiliar topic. Students were randomly assigned to an AI-permitted or an AI-forbidden condition. Students in the AI-permitted condition scored higher than those in the AI-forbidden condition on an immediate knowledge test (during which there was no AI access), although their essays were not judged better. One week later, all students were tested again and wrote another essay; no AI was permitted. At this follow-up assessment, the test advantage of those in the original AI-permitted condition was again observed, and their essays were judged superior to those of students in the AI-forbidden condition. However, Contractor and Reyes observed that the benefits of AI use were greater for augmentation users than for automation users, with this classification determined from an analysis of the logs of students\u2019 interactions with AI in the first session.\r\n\r\nThese studies suggest that when properly used, AI may benefit learning, but may also be used to complete tasks bypassing learning. AI assistance often improves immediate task performance, but these performance gains are not necessarily accompanied by learning.\r\n\r\nThe reviewed studies examined learning only over a short term (a couple of hours to one week). Bastani et al. (2025) noted that studying longer-term outcomes should be an emphasis of future research.\r\n\r\nLodge and Loble (2026) present an annotated bibliography that includes summaries of experimental investigations of the effects of AI on learning; Contractor and Reyes (2026) summarized and conducted a meta-analysis of such studies.\r\n

Implications<\/h3>\r\nThe effects on learning of AI use depend on how students use it. The ubiquitous accessibility of AI and its utility for carrying out both lower-level and higher-level academic tasks makes it essential that instructional faculty carefully determine exactly what students need to know and design procedures for assessing whether students have achieved that learning (Chirikov et al., 2026; Cooper, 2026; Lodge & Loble, 2026).\r\n

3. Student Experience and Policy Consistency<\/h2>\r\n

Clarity and Transparency in AI Expectations<\/h3>\r\nFragmented AI policies across courses can create confusion for students regarding what constitutes appropriate, ethical, and academically acceptable AI use. Research examining students\u2019 perceptions of generative AI governance in higher education further underscores the need for clear institutional guidelines, ethical guidance, and AI competency development to help students navigate the appropriate use of AI in academic settings (Barus et al., 2025). These findings highlight the importance of establishing transparent expectations that help students understand not only when AI use is appropriate, but also how to engage with these tools responsibly and ethically across different academic contexts.\r\n\r\nWhen expectations vary significantly from one instructor or course to another, students may unintentionally violate academic integrity policies or become uncertain about when and how AI can appropriately support their learning. Institutions should\u00a0 prioritize clear, transparent, and developmentally informed AI policies that help students understand not only what uses of AI are permitted, but also why certain learning activities may require independent thinking and engagement.\r\n

AI Use and Cognitive Development<\/h3>\r\nAn important consideration that extends beyond academic integrity is the potential impact of AI on students\u2019 cognitive development. College is a critical period for developing higher-order thinking skills, including critical thinking, problem-solving, decision-making, metacognition, information evaluation, and the ability to construct and defend an independent argument. As generative AI becomes increasingly integrated into learning, scholars have emphasized the importance of considering how its use influences students\u2019 cognitive engagement and the learning processes necessary for developing knowledge and higher-order thinking skills (Bauer et al., 2025). If generative AI is routinely used to perform cognitive tasks that students would otherwise complete themselves such as generating ideas, synthesizing readings, constructing arguments, or solving problems students may have fewer opportunities to practice and strengthen these skills.\r\n\r\nThis concern may be especially important when considering students\u2019 developmental trajectories. Emerging adulthood is a period in which executive functioning, self-regulation, judgment, and complex decision-making continue to develop. Consequently, higher education should consider not simply whether students are using AI, but how AI use interacts with the developmental processes that occur during the college years. Overreliance on AI may encourage cognitive offloading, in which students transfer portions of the thinking process to technology rather than engaging deeply with the material themselves. Farhat (2026) found that overreliance on generative AI may be associated with reduced critical thinking, weaker memory retention, and increased cognitive dependence, highlighting potential concerns when AI substitutes for active cognitive engagement. At the same time, Farhat\u2019s findings suggest that AI can be used more constructively as a scaffold for creativity and reflective learning. Appropriately structured AI use may therefore support students in questioning information, comparing perspectives, receiving formative feedback, and refining their thinking rather than replacing the cognitive processes essential to learning.\r\n

Consistency Without Uniformity<\/h3>\r\nTherefore, institutional AI policies should move beyond a binary framework of \u201cAI allowed\u201d versus \u201cAI prohibited.\u201d A more developmentally responsive approach would identify which cognitive processes students should first demonstrate independently and when AI can subsequently be introduced as a tool to extend or enhance learning. Importantly, AI expectations may need to vary not only across courses and disciplines but also across assignments within the same course. Because assignments are designed to assess different learning objectives and cognitive skills, the appropriate level of AI assistance may differ accordingly. For example, an assignment intended to assess a student\u2019s ability to independently construct an argument may appropriately restrict AI use, whereas another assignment may invite students to use AI to critique their reasoning, identify alternative perspectives, or receive formative feedback. In this way, AI use is intentionally aligned with learning objectives rather than governed by a single course-wide designation.\r\n\r\nConsistency, therefore, should not necessarily mean uniformity in AI use across courses or assignments. Instead, institutions can establish common principles for responsible AI use while allowing faculty to develop discipline, course, and assignment-specific expectations based on intended learning outcomes. Wang et al. (2024) found that many universities are adopting this type of contextualized approach, with institutions providing overarching guidance and resources while allowing instructors to make decisions about GenAI use within their specific teaching contexts. The authors further emphasize aligning GenAI use with learning objectives and recommend discipline-specific policies and guidelines rather than a universal approach to AI use.\r\n\r\nThis approach is also reflected in instructional guidance from the University of Georgia\u2019s Center for Teaching and Learning. Vanderveen and Poproski (2026) encourage faculty to consider learning goals when establishing expectations for generative AI and recognize that appropriate AI use may differ across learning activities and assignments. Clearly communicating when AI is permitted or prohibited, expectations for disclosure or citation, and the educational rationale behind these decisions can reduce ambiguity while supporting students\u2019 development of AI literacy.\r\n\r\nImportantly, explaining the rationale for AI expectations can also help students understand that restrictions are not simply punitive or arbitrary but may be intentionally designed to preserve opportunities to develop particular knowledge and cognitive skills. Ultimately, the goal should extend beyond preventing academic misconduct to ensuring that students graduate with the intellectual independence, critical-thinking abilities, ethical judgment, and AI literacy necessary to navigate an increasingly AI-mediated academic and professional environment.\r\n

4. Demystifying the Use of Artificial Intelligence in Higher Education and the Workplace<\/h2>\r\nAs artificial intelligence becomes increasingly integrated into professional practice, employers increasingly expect college graduates to possess foundational AI literacy and the ability to use AI tools effectively and responsibly. In addition to technical proficiency, employers increasingly value graduates who can critically evaluate AI-generated information, verify its accuracy, and apply disciplinary knowledge and professional judgment when using AI to support decision-making. Findings from the Fall 2025 CSU LIFT Team Survey (Elosh, Kumar, & Sukhoy), which included responses from College of Business alumni, Washkewicz College of Engineering alumni, and professional contacts, revealed a notable disconnect between employer expectations and employer perceptions of recent graduates' preparedness.\r\n\r\nAmong the 63 respondents, 70% indicated that AI-related knowledge and skills are expected of new graduates. However, only 10% expressed confidence that recent graduates possess adequate AI competencies. <\/em>This disparity highlights a significant opportunity for higher education institutions to better prepare students for an AI-enabled workforce.\r\n\r\nDemystifying AI within higher education requires attention to two interrelated dimensions. First, students must learn to distinguish between AI as a form of instrumental help-seeking that supports learning and productivity, and AI as a form of executive help-seeking that substitutes for the learner's cognitive effort (Aguilar, 2025; USC, 2025). Second, institutions must intentionally assess student knowledge development both with and without AI support in order to understand how AI influences learning outcomes, critical thinking, and disciplinary mastery.\r\n\r\nThese objectives can be advanced through a range of faculty-driven instructional strategies that are applicable across academic disciplines.\r\n

Scaffolded AI Integration in Major Projects<\/h3>\r\nComplex assignments can be intentionally scaffolded to include varying levels of AI use, allowing faculty to define when and how AI tools may be utilized. Appropriate applications may include brainstorming, topic development, outlining, exploratory research, and refinement of arguments. Such structures encourage students to leverage AI as a cognitive support tool while maintaining responsibility for verification, analysis, evaluation, and final decision-making. Students should also document how AI was used and explain how they validated AI-generated information within the context of the assignment.\r\n

Collaborative Exploration of AI Capabilities and Limitations<\/h3>\r\nClassroom discussions and group activities can provide opportunities for students to engage directly with evolving AI technologies. These experiences may include prompt-design exercises, comparative analyses of outputs generated by different AI platforms, and evaluation of factual accuracy, bias, and hallucinations. Such activities encourage students to become critical consumers of AI-generated content rather than passive recipients.\r\n

Discipline-Specific AI Case Analysis<\/h3>\r\nStudents can investigate contemporary organizational, business, healthcare, engineering, and public-sector applications of AI. By examining current use cases, emerging capabilities, regulatory developments, and ethical considerations, students gain a more nuanced understanding of AI's role across industries. These analyses may be conducted individually or collaboratively and shared through written submissions, presentations, or facilitated class discussions. Faculty should encourage students to evaluate not only the opportunities presented by AI but also its limitations, uncertainties, and potential risks within their discipline.\r\n

Developing AI Evaluation Literacy<\/h3>\r\nGiven the limitations of current AI-detection technologies and the prevalence of false positives, students should be educated on the strengths and weaknesses of AI-evaluation tools. Rather than treating AI detectors as definitive sources of truth, students should learn to identify writing patterns, assess transparency of AI use, and evaluate the quality and originality of their own work. Developing this literacy encourages responsible use while supporting academic integrity.\r\n

AI Transparency and Reflective Practice<\/h3>\r\nMajor assignments can incorporate AI disclosure statements\/declarations that require students to document the nature and extent of AI use. These reflections may include the tools utilized, specific tasks performed with AI assistance, perceived benefits, limitations encountered, and lessons learned throughout the process. Such practices promote transparency while encouraging metacognitive reflection about learning. They also help faculty distinguish appropriate AI-assisted learning from excessive dependence on AI-generated work.\r\n

Independent Inquiry and Knowledge Sharing<\/h3>\r\nStudents may investigate emerging AI applications, compare multiple AI tools, and evaluate how different approaches influence the quality, reliability, and efficiency of their work. Sharing findings with peers creates a collaborative learning environment while helping students recognize the evolving relationship between AI technologies and their future careers.\r\n

Faculty and Industry Partnerships<\/h3>\r\nInviting faculty members from multiple disciplines, alongside industry practitioners, can expose students to diverse perspectives on AI implementation. Guest speakers can help connect technical developments, ethical considerations, and workplace expectations to students' particular fields of study. Industry engagement helps students understand how AI is currently being used in professional practice and reinforces that successful AI adoption requires domain expertise, engineering judgment, ethical decision-making, and effective communication.\r\n

Alumni Engagement<\/h3>\r\nRecent graduates offer a particularly valuable perspective because they can speak directly to the transition from college to the workforce. Alumni can provide practical insights regarding AI adoption in professional settings, employer expectations, and the competencies students should cultivate to remain competitive and effective in their careers.\r\n

Leveraging Institutional Support Resources<\/h3>\r\nInstitutions can further strengthen AI literacy by encouraging students to engage with existing campus resources such as writing centers, libraries, tutoring services, and dedicated AI learning labs. These resources can help students develop both technical proficiency and critical evaluation skills.\r\n

In Conclusion<\/h3>\r\nCollectively, these instructional approaches emphasize critical evaluation, reflective practice, and human judgment. They encourage students to engage with AI as a tool that augments learning and professional practice rather than one that replaces critical thinking, disciplinary expertise, or human judgment. In doing so, higher education can promote intentional forms of AI use while discouraging excessive dependence on AI for cognitive tasks that are central to student development. Developing AI literacy in this way is also an important component of workforce readiness, as students must learn not only how to use AI tools effectively, but also when their use is appropriate, how to critically evaluate AI-generated information, and when professional judgment and human expertise must take precedence.\r\n\r\nUltimately, these practices can help narrow the gap between employer expectations and perceived graduate preparedness. By intentionally integrating AI literacy into teaching and learning, colleges and universities can better prepare students for an increasingly AI-enabled workplace. In this context, AI literacy extends beyond technical proficiency to include the critical thinking, ethical decision-making, and professional judgment needed to evaluate AI-generated information and make responsible decisions within professional settings. Positioning AI literacy as a component of workforce readiness can help ensure that graduates are prepared not simply to use emerging technologies, but to exercise the disciplinary knowledge, intellectual independence, and professional judgment necessary to determine how and when those technologies should be used.\r\n

5. AI Literacy Outcomes for Students<\/h2>\r\nAI literacy should be understood as a multidimensional educational outcome, not simply the ability to access an AI tool or write an effective prompt. It includes sufficient knowledge of how AI systems use data and generate outputs; the practical ability to select, operate, and integrate appropriate tools; the capacity to evaluate their performance and limitations; and an understanding of their ethical, legal, and societal implications. Hackl et al. (2026) organize these capacities into seven interconnected dimensions: technical knowledge, application proficiency, critical thinking, ethical reasoning, social-impact understanding, integration skills, and legal and regulatory knowledge. Similarly, Kang and Park\u2019s (2026) validated scale for higher education students identifies four related dimensions: fundamental knowledge, practical application, performance evaluation, and impact assessment. Although Zhou et al. (2025) distinguish AI literacy, or what individuals know, from AI competency, or how effectively and reflectively they apply that knowledge, higher education should cultivate both.\r\n\r\nAn AI-literate graduate should therefore be able to explain, at an appropriate level, how AI systems are trained and why their outputs can be incomplete, inaccurate, biased, or fabricated. Students should be able to identify when AI is suitable for a task, formulate and refine productive queries, and integrate its outputs with disciplinary knowledge without surrendering human judgment. They should know how to verify claims against credible evidence, compare alternative outputs, recognize uncertainty and embedded assumptions, and determine whether AI has strengthened or displaced their own reasoning. AI literacy also requires students to protect personal and confidential information, respect intellectual property, document and disclose their use of AI, and accept responsibility for the resulting work. Finally, students should be prepared to evaluate AI\u2019s broader consequences for fairness, access, employment, human relationships, democratic life, and the environment. Because the technology and its governing rules will continue to change, students must also learn to assess their own capabilities, recognize overreliance, and update their knowledge over time (Zhou et al., 2025). These outcomes should be taught and assessed throughout the curriculum rather than relegated to academic-integrity policies or optional technical training.\r\n

6. Conclusion and Resources<\/h2>\r\nResponsible AI integration is fundamentally a pedagogical challenge that higher education institutions must address proactively and intentionally. As AI becomes increasingly embedded across professional sectors, employers will expect college graduates to possess both AI proficiency and the critical-thinking skills necessary to evaluate, apply, and question AI-generated information. Therefore, the development of AI literacy should not occur at the expense of the cognitive skills and disciplinary knowledge that students need to exercise independent and professional judgment.\r\n\r\nHigher education institutions have an important role in creating learning environments in which students develop AI proficiency alongside critical thinking, ethical decision-making, and responsible technology use. Rather than focusing solely on whether students should or should not use AI, educators should consider how AI can be intentionally integrated into teaching and learning in ways that support specific learning objectives while preserving opportunities for students to engage in independent thinking, problem-solving, and knowledge development.\r\n\r\nCollaboration among faculty, students, and institutions is essential to this process. Establishing clear expectations for responsible and ethical AI use, developing students\u2019 technical and critical AI literacy, and increasing awareness of how AI is being used across professional sectors can help demystify emerging technologies while preparing students to use them effectively and responsibly. Ultimately, higher education must prepare graduates not simply to use AI, but to determine when and how it should be used, critically evaluate its outputs and limitations, and exercise the human judgment necessary to make ethical and informed decisions in an increasingly AI-mediated workforce.\r\n

Resources for Responsible AI Integration<\/h3>\r\nSeveral resources can support institutions and faculty as they navigate the responsible integration of AI into teaching and learning. The EDUCAUSE AI Literacy in Teaching and Learning: A Durable Framework for Higher Education<\/em> provides a comprehensive framework for developing AI literacy among students, faculty, and staff. The framework emphasizes understanding AI fundamentals, critically evaluating AI tools and outputs, recognizing limitations and potential biases, and using AI effectively and ethically in academic and professional contexts (Kassorla et al., 2024).\r\n\r\nEDUCAUSE also maintains a Teaching with Artificial Intelligence<\/em> resource collection that brings together practical materials related to classroom AI use, academic integrity, course design, AI literacy, and policy development (EDUCAUSE, n.d.). These resources can assist faculty and institutional leaders in translating broader principles of responsible AI use into teaching and learning practices.\r\n\r\nAdditionally, the University of Georgia Center for Teaching and Learning provides practical guidance for developing generative AI policies at the course and assignment levels. Vanderveen and Poproski (2026.) emphasize the importance of clearly communicating AI expectations and their rationale while recognizing that different learning goals may warrant different levels of AI use within the same course. This approach provides faculty with a practical model for aligning AI expectations with learning objectives rather than relying on a single universal policy for all learning activities.\r\n\r\nTogether, these resources offer faculty and institutions practical starting points for developing AI literacy, establishing transparent policies, aligning AI use with learning objectives, and maintaining human judgment and critical thinking at the center of teaching and learning.\r\n

7. AI Statement<\/h2>\r\nSection 1: AI Statement: Researched and drafted by a human being without AI assistance; ChatGPT assisted with editing and revision.\r\n\r\nSection 2: AI Statement: Researched and drafted by a human being without AI assistance; ChatGPT assisted with editing and revision.\r\n\r\nSection 3: Researched and drafted by a human being without AI assistance; ChatGPT assisted with editing and revision.\r\n\r\nSection 4: AI Statement: Researched and drafted by a human being without AI assistance; CoPilot assisted with editing and revision.\r\n\r\nSection 5: AI Statement: Researched and drafted by a human being without AI assistance; ChatGPT assisted with editing and revision.\r\n

References<\/h2>\r\n

Aguilar, S. J., Nye, B., Swartout, W. R., Macias, A., Xing, Y., & Xiu, R. (2025, August). Fostering critical thinking in the age of AI<\/em>. <\/a>https:\/\/doi.org\/10.35542\/osf.io\/wr6n3_v2<\/a><\/p>\r\n

Barus, O. P., Hidayanto, A. N., Handri, E. Y., Sensuse, D. I., & Yaiprasert, C. (2025). Shaping generative AI governance in higher education: Insights from student perception. International Journal of Educational Research Open, 8<\/em>, Article 100452. https:\/\/doi.org\/10.1016\/j.ijedro.2025.100452<\/a><\/p>\r\n

Bassner, P., Lenk-Ostendorf, B., Beinstingel, R., Wasner, T., & Krusche, S. (2026). Less stress, better scores, same learning: The dissociation of performance and learning in AI-supported programming education. Computers and Education: Artificial Intelligence, 10<\/em>, Article 100537. <\/a>https:\/\/doi.org\/10.1016\/j.caeai.2025.100537<\/a><\/p>\r\n

Bastani, H., Bastani, O., Sungu, A., Ge, H., Kabakci, O., & Mariman, R. (2025). Generative AI without guardrails can harm learning: Evidence from high school mathematics. Proceedings of the National Academy of Sciences, 122<\/em>(26), Article e2422633122. <\/a>https:\/\/doi.org\/10.1073\/pnas.2422633122<\/a><\/p>\r\n

Bauer, E., Greiff, S., Graesser, A. C., Scheiter, K., & Sailer, M. (2025). Looking beyond the hype: Understanding the effects of AI on learning. Educational Psychology Review, 37<\/em>, Article 45. <\/a>https:\/\/doi.org\/10.1007\/s10648-025-10020-8<\/a><\/p>\r\n

Chirkov, I., Smirnov, I., & Kizilcec, R. F. (2026). Generative AI use and misuse call for assessment reform in higher education. Science, 392<\/em>(6800), 818\u2013820. <\/a>https:\/\/doi.org\/10.1126\/science.aec5115<\/a><\/p>\r\n

Contractor, Z., & Reyes, G. (2026). Experimental evidence on the learning impact of generative AI<\/em> (IZA Discussion Paper No. 18792). IZA Institute of Labor Economics. <\/a>https:\/\/www.iza.org\/publications\/dp\/18792<\/a><\/p>\r\n

Cooper, M. M. (2026). A preliminary set of principles to support learning in the context of generative AI. Journal of Chemical Education, 103<\/em>, 36\u201342. <\/a>https:\/\/doi.org\/10.1021\/acs.jchemed.5c01413<\/a><\/p>\r\n

EDUCAUSE. (n.d.). Teaching with artificial intelligence<\/em>. Teaching with Artificial Intelligence<\/a><\/p>\r\n

Elosh, E., Kumar, S., & Sukhoy, A. (2025). Integrating AI at CSU: Exploring the current state and future opportunities with our faculty, students and alumni.<\/em><\/p>\r\n

Farhat, Z. (2026). When AI thinks for us: Unveiling the cognitive and social toll of ChatGPT over-reliance. Applied Cognitive Psychology, 40<\/em>, Article e70205. <\/a>https:\/\/doi.org\/10.1002\/acp.70205<\/a><\/p>\r\n

Hackl, V., M\u00fcller, A. E., & Sailer, M. (2026). The AI literacy heptagon: A structured approach to AI literacy in higher education. Computers and Education: Artificial Intelligence, 10<\/em>, Article 100540. <\/a>https:\/\/doi.org\/10.1016\/j.caeai.2026.100540<\/a><\/p>\r\n

Hon, K. (2026). Generative AI in higher education: A systematic review of its effects on learning outcomes and academic performance. Journal of Educational Technology Systems, 54<\/em>(3), 537\u2013560. <\/a>https:\/\/doi.org\/10.1177\/00472395251400089<\/a><\/p>\r\n

Jose, B., Cherian, J., Verghis, A. M., Varghise, S. M., S, M., & Joseph, S. (2025). The cognitive paradox of AI in education: Between enhancement and erosion. Frontiers in Psychology, 16<\/em>, Article 1550621. <\/a>https:\/\/doi.org\/10.3389\/fpsyg.2025.1550621<\/a><\/p>\r\n

Kang, K. Y., & Park, J.-H. (2026). Development and validation of a multidimensional AI literacy scale for higher education students: A mixed-method study. Journal of Librarianship and Information Science<\/em>. Advance online publication. <\/a>https:\/\/doi.org\/10.1177\/09610006261425235<\/a><\/p>\r\n

Kassorla, M., Georgieva, M., & Papini, A. (2024). AI literacy in teaching and learning: A durable framework for higher education<\/em>. EDUCAUSE. AI Literacy in Teaching and Learning<\/a><\/p>\r\n

Lodge, J. M., & Loble, L. (2026). Artificial intelligence, cognitive offloading and implications for education<\/em>. University of Technology Sydney. <\/a>https:\/\/doi.org\/10.71741\/4pyxmbnjaq.31302475<\/a><\/p>\r\n

Pallant, J. L., Blijlevens, J., Campbell, A., & Jopp, R. (2026). Mastering knowledge: The impact of generative AI on student learning outcomes. Studies in Higher Education, 51<\/em>(4), 714\u2013735. <\/a>https:\/\/doi.org\/10.1080\/03075079.2025.2487570<\/a><\/p>\r\n

Slade, J. J. (2025). How to help your students use AI without losing the learning<\/em>. American Psychological Association. <\/a>https:\/\/www.apa.org\/ed\/precollege\/psychology-teacher-network\/introductory-psychology\/learning-artificial-intelligence<\/a><\/p>\r\n

Vanderveen, T., & Poproski, R. (2026). Generative AI policies for teaching & learning<\/em>. University of Georgia Center for Teaching and Learning. Generative AI Policies for Teaching & Learning<\/a><\/p>\r\n

Wang, H., Dang, A., Wu, Z., & Mac, S. (2024). Generative AI in higher education: Seeing ChatGPT through universities' policies, resources, and guidelines. Computers and Education: Artificial Intelligence, 7<\/em>, Article 100326. https:\/\/doi.org\/10.1016\/j.caeai.2024.100326<\/a><\/p>\r\n

Zhou, X., Li, Y., Chai, C. S., & Chiu, T. K. F. (2025). Defining, enhancing, and assessing artificial intelligence literacy and competency in K\u201312 education from a systematic review. Interactive Learning Environments, 33<\/em>(10), 5766\u20135788. <\/a>https:\/\/doi.org\/10.1080\/10494820.2025.2487538<\/a><\/p>\r\n ","rendered":"

Navid Goudarzi, Dakota King-White, Fred Smith, Alex Sukhoy, Adam Voight<\/p>\n

Abstract<\/h2>\n

Artificial intelligence (AI) has become an embedded feature of contemporary academic life, shaping how students research, write, study, and prepare for professional environments. This paper articulates an instructional framework for integrating AI into higher education that prioritizes learning outcomes, ethical reasoning, and human judgment. Drawing on Cleveland State University instructional practice, the framework distinguishes between instrumental and executive uses of AI and argues for faculty-guided structures that support transparency, critical evaluation, and workforce readiness.<\/p>\n

1. Introduction on Research on how AI is Impacting Student Learning<\/h2>\n

Generative AI has rapidly become embedded in how college students research, write, solve problems, obtain feedback, and prepare for professional work. Because students will encounter these tools both inside and outside the university, treating AI primarily as a new form of academic misconduct is no longer an adequate institutional response. The more consequential question is what students must learn to use AI productively, ethically, and critically. Higher education must prepare students not merely to generate competent outputs, but to exercise judgment about when AI is appropriate, evaluate the accuracy and limitations of its responses, recognize bias, protect sensitive information, and remain accountable for the work they produce. These capabilities will increasingly shape graduates\u2019 readiness to compete in the workplace and contribute responsibly to their professions and communities.<\/p>\n

Emerging research suggests that AI\u2019s effects on learning depend substantially on how it is used. Hon\u2019s (2026) systematic review of 41 studies found that generative AI can increase engagement, provide personalized feedback and explanations, improve efficiency, and support academic performance. At the same time, the review identified risks of cognitive offloading, superficial engagement, overreliance, and diminished critical thinking when students accept AI-generated content without scrutiny. Pallant et al.\u2019s (2026) analysis of 192 student reflections similarly found that students who used AI to construct and augment their own knowledge demonstrated higher grades, applied knowledge, critical thinking, and learning autonomy. Students who used it procedurally or simply reproduced its outputs demonstrated poorer outcomes. Although this evidence remains developing, it indicates that AI is neither inherently beneficial nor inherently harmful to learning. Its educational value depends on whether students possess the skills and dispositions needed to use it as a resource for thinking rather than as a substitute for thinking.<\/p>\n

2. Instrumental vs. Executive AI Use<\/h2>\n

Instrumental AI use supports learning, while executive AI use replaces cognitive labor. Effective pedagogy bridges this gap by centering evaluation and judgment with the student.<\/p>\n

Use of generative AI tools by students may have both positive and negative effects on student learning. Use of AI to automate routine or lower-level tasks (e.g., checking grammar; doing calculations) has the potential to free time and cognitive resources to engage in higher-level tasks, increasing learning. However, if AI is used for those higher-level tasks, the potential benefit for learning may be lost.<\/p>\n

In this section, we distinguish two ways in which generative AI tools might be used by students; describe the cognitive consequences of those uses; and provide some examples of the types of research conducted to reach these conclusions.<\/p>\n

Ways Students May Use Generative AI Tools<\/h3>\n

Various researchers have distinguished between use of generative AI to support learning and use of AI to replace learning (e.g., Slade, 2025), and there is evidence that students use AI in each of these ways. Student learning may be supported by using AI tutors, obtaining feedback from AI systems on the quality and completeness of an essay, or using AI to corroborate the solution to a problem (see, e.g., Jose et al., 2025). These uses have been characterized as instrumental (Aguilar et al., 2025); Contractor and Reyes (2026) labeled students who use AI this way as augmentation users. When students engage in such uses of AI, they engage in the cognitive processes that support learning.<\/p>\n

When students use AI to write an essay or solve problems with little or no cognitive engagement, the opportunity for learning is reduced. Aguilar et al. characterized such uses of AI as executive uses; Bauer et al. (2025) call this an inversion effect; Contractor and Reyes (2026) labeled students who use AI this way as automation users. Surveys show that students report engaging in both types of AI use (e.g., Aguilar et al.), as do content analyses of logs of student interactions with generative AI tools during experimental studies in which use of AI tools is permitted (e.g., Contractor & Reyes, 2026). Because durable learning depends on students generating explanations, retrieving knowledge, organizing information, and solving problems themselves, replacing these processes with AI reduces opportunities for learning.<\/p>\n

Instrumental use of AI represents selective cognitive offloading: routine operations are delegated while students retain responsibility for the cognitive processes central to learning. Executive use of AI involves offloading those very processes responsible for constructing knowledge.<\/p>\n

Consequences of AI Use<\/h3>\n

Instrumental use of AI has the potential to promote learning because while using AI tools in these ways, students should be engaged in the cognitive processes that underlie learning while using AI to support their work. In contrast, because executive use essentially outsources those aspects of the task that engage those cognitive processes, opportunities for learning are reduced.<\/p>\n

In numerous studies, students have been asked to report on their use (and misuse) of AI, their engagement with their academic tasks, and their perceptions of the effects of AI on their learning (e.g., Aguilar et al., 2025; Chirkov et al., 2026; Farhat, 2026). Although \u201cthere is no robust body of work on actual learning\u201d (Cooper, 2026, p. 37), a body of work is developing in which researchers are investigating effects on learning of using various AI tools. Although many of these studies are methodologically weak (see, e.g., Bauer et al., 2025), the several studies described here, although imperfect, exemplify rigorous experimental investigations of the effects on learning of AI use and illustrate variability in findings concerning learning.<\/p>\n

Bastani et al. (2025) concluded that access to generative AI during practice of math problems negatively affected learning. They randomly assigned math classrooms with close to 1000 Turkish high school students to three experimental conditions that differed according to a resource that was available while students practiced a math concept for 90 minutes. The conditions were access to ChatGPT, access to a specialized ChatGPT tutor that would provide hints and guidance and permit checking of answers, and a no-AI control. The tutor condition constrained AI use (encouraging instrumental AI use), whereas unrestricted access to ChatGPT permitted executive AI use. Following the practice session, students took a closed-book exam. Although students in the ChatGPT conditions performed better on the practice problems than those in the no-AI control condition, these benefits did not carry over to performance on the exam. On the exam, performance of students in the no-AI control condition exceeded that of those in the ChatGPT condition and did not differ significantly from those in the ChatGPT tutor condition. Bassner et al. (2026) conducted a similar study using a programming task with undergraduate- and master\u2019s-level management students, and reached similar conclusions: Although students in the AI conditions outperformed those in the no-AI control condition on the task, this advantage was not reflected in better performance on the measures used to assess learning. This replicated an important pattern observed by Bastani et al.: AI improved immediate performance without corresponding improvements in learning.<\/p>\n

Contractor and Reyes (2026) tasked students at an elite college with learning about and writing an essay on an unfamiliar topic. Students were randomly assigned to an AI-permitted or an AI-forbidden condition. Students in the AI-permitted condition scored higher than those in the AI-forbidden condition on an immediate knowledge test (during which there was no AI access), although their essays were not judged better. One week later, all students were tested again and wrote another essay; no AI was permitted. At this follow-up assessment, the test advantage of those in the original AI-permitted condition was again observed, and their essays were judged superior to those of students in the AI-forbidden condition. However, Contractor and Reyes observed that the benefits of AI use were greater for augmentation users than for automation users, with this classification determined from an analysis of the logs of students\u2019 interactions with AI in the first session.<\/p>\n

These studies suggest that when properly used, AI may benefit learning, but may also be used to complete tasks bypassing learning. AI assistance often improves immediate task performance, but these performance gains are not necessarily accompanied by learning.<\/p>\n

The reviewed studies examined learning only over a short term (a couple of hours to one week). Bastani et al. (2025) noted that studying longer-term outcomes should be an emphasis of future research.<\/p>\n

Lodge and Loble (2026) present an annotated bibliography that includes summaries of experimental investigations of the effects of AI on learning; Contractor and Reyes (2026) summarized and conducted a meta-analysis of such studies.<\/p>\n

Implications<\/h3>\n

The effects on learning of AI use depend on how students use it. The ubiquitous accessibility of AI and its utility for carrying out both lower-level and higher-level academic tasks makes it essential that instructional faculty carefully determine exactly what students need to know and design procedures for assessing whether students have achieved that learning (Chirikov et al., 2026; Cooper, 2026; Lodge & Loble, 2026).<\/p>\n

3. Student Experience and Policy Consistency<\/h2>\n

Clarity and Transparency in AI Expectations<\/h3>\n

Fragmented AI policies across courses can create confusion for students regarding what constitutes appropriate, ethical, and academically acceptable AI use. Research examining students\u2019 perceptions of generative AI governance in higher education further underscores the need for clear institutional guidelines, ethical guidance, and AI competency development to help students navigate the appropriate use of AI in academic settings (Barus et al., 2025). These findings highlight the importance of establishing transparent expectations that help students understand not only when AI use is appropriate, but also how to engage with these tools responsibly and ethically across different academic contexts.<\/p>\n

When expectations vary significantly from one instructor or course to another, students may unintentionally violate academic integrity policies or become uncertain about when and how AI can appropriately support their learning. Institutions should\u00a0 prioritize clear, transparent, and developmentally informed AI policies that help students understand not only what uses of AI are permitted, but also why certain learning activities may require independent thinking and engagement.<\/p>\n

AI Use and Cognitive Development<\/h3>\n

An important consideration that extends beyond academic integrity is the potential impact of AI on students\u2019 cognitive development. College is a critical period for developing higher-order thinking skills, including critical thinking, problem-solving, decision-making, metacognition, information evaluation, and the ability to construct and defend an independent argument. As generative AI becomes increasingly integrated into learning, scholars have emphasized the importance of considering how its use influences students\u2019 cognitive engagement and the learning processes necessary for developing knowledge and higher-order thinking skills (Bauer et al., 2025). If generative AI is routinely used to perform cognitive tasks that students would otherwise complete themselves such as generating ideas, synthesizing readings, constructing arguments, or solving problems students may have fewer opportunities to practice and strengthen these skills.<\/p>\n

This concern may be especially important when considering students\u2019 developmental trajectories. Emerging adulthood is a period in which executive functioning, self-regulation, judgment, and complex decision-making continue to develop. Consequently, higher education should consider not simply whether students are using AI, but how AI use interacts with the developmental processes that occur during the college years. Overreliance on AI may encourage cognitive offloading, in which students transfer portions of the thinking process to technology rather than engaging deeply with the material themselves. Farhat (2026) found that overreliance on generative AI may be associated with reduced critical thinking, weaker memory retention, and increased cognitive dependence, highlighting potential concerns when AI substitutes for active cognitive engagement. At the same time, Farhat\u2019s findings suggest that AI can be used more constructively as a scaffold for creativity and reflective learning. Appropriately structured AI use may therefore support students in questioning information, comparing perspectives, receiving formative feedback, and refining their thinking rather than replacing the cognitive processes essential to learning.<\/p>\n

Consistency Without Uniformity<\/h3>\n

Therefore, institutional AI policies should move beyond a binary framework of \u201cAI allowed\u201d versus \u201cAI prohibited.\u201d A more developmentally responsive approach would identify which cognitive processes students should first demonstrate independently and when AI can subsequently be introduced as a tool to extend or enhance learning. Importantly, AI expectations may need to vary not only across courses and disciplines but also across assignments within the same course. Because assignments are designed to assess different learning objectives and cognitive skills, the appropriate level of AI assistance may differ accordingly. For example, an assignment intended to assess a student\u2019s ability to independently construct an argument may appropriately restrict AI use, whereas another assignment may invite students to use AI to critique their reasoning, identify alternative perspectives, or receive formative feedback. In this way, AI use is intentionally aligned with learning objectives rather than governed by a single course-wide designation.<\/p>\n

Consistency, therefore, should not necessarily mean uniformity in AI use across courses or assignments. Instead, institutions can establish common principles for responsible AI use while allowing faculty to develop discipline, course, and assignment-specific expectations based on intended learning outcomes. Wang et al. (2024) found that many universities are adopting this type of contextualized approach, with institutions providing overarching guidance and resources while allowing instructors to make decisions about GenAI use within their specific teaching contexts. The authors further emphasize aligning GenAI use with learning objectives and recommend discipline-specific policies and guidelines rather than a universal approach to AI use.<\/p>\n

This approach is also reflected in instructional guidance from the University of Georgia\u2019s Center for Teaching and Learning. Vanderveen and Poproski (2026) encourage faculty to consider learning goals when establishing expectations for generative AI and recognize that appropriate AI use may differ across learning activities and assignments. Clearly communicating when AI is permitted or prohibited, expectations for disclosure or citation, and the educational rationale behind these decisions can reduce ambiguity while supporting students\u2019 development of AI literacy.<\/p>\n

Importantly, explaining the rationale for AI expectations can also help students understand that restrictions are not simply punitive or arbitrary but may be intentionally designed to preserve opportunities to develop particular knowledge and cognitive skills. Ultimately, the goal should extend beyond preventing academic misconduct to ensuring that students graduate with the intellectual independence, critical-thinking abilities, ethical judgment, and AI literacy necessary to navigate an increasingly AI-mediated academic and professional environment.<\/p>\n

4. Demystifying the Use of Artificial Intelligence in Higher Education and the Workplace<\/h2>\n

As artificial intelligence becomes increasingly integrated into professional practice, employers increasingly expect college graduates to possess foundational AI literacy and the ability to use AI tools effectively and responsibly. In addition to technical proficiency, employers increasingly value graduates who can critically evaluate AI-generated information, verify its accuracy, and apply disciplinary knowledge and professional judgment when using AI to support decision-making. Findings from the Fall 2025 CSU LIFT Team Survey (Elosh, Kumar, & Sukhoy), which included responses from College of Business alumni, Washkewicz College of Engineering alumni, and professional contacts, revealed a notable disconnect between employer expectations and employer perceptions of recent graduates’ preparedness.<\/p>\n

Among the 63 respondents, 70% indicated that AI-related knowledge and skills are expected of new graduates. However, only 10% expressed confidence that recent graduates possess adequate AI competencies. <\/em>This disparity highlights a significant opportunity for higher education institutions to better prepare students for an AI-enabled workforce.<\/p>\n

Demystifying AI within higher education requires attention to two interrelated dimensions. First, students must learn to distinguish between AI as a form of instrumental help-seeking that supports learning and productivity, and AI as a form of executive help-seeking that substitutes for the learner’s cognitive effort (Aguilar, 2025; USC, 2025). Second, institutions must intentionally assess student knowledge development both with and without AI support in order to understand how AI influences learning outcomes, critical thinking, and disciplinary mastery.<\/p>\n

These objectives can be advanced through a range of faculty-driven instructional strategies that are applicable across academic disciplines.<\/p>\n

Scaffolded AI Integration in Major Projects<\/h3>\n

Complex assignments can be intentionally scaffolded to include varying levels of AI use, allowing faculty to define when and how AI tools may be utilized. Appropriate applications may include brainstorming, topic development, outlining, exploratory research, and refinement of arguments. Such structures encourage students to leverage AI as a cognitive support tool while maintaining responsibility for verification, analysis, evaluation, and final decision-making. Students should also document how AI was used and explain how they validated AI-generated information within the context of the assignment.<\/p>\n

Collaborative Exploration of AI Capabilities and Limitations<\/h3>\n

Classroom discussions and group activities can provide opportunities for students to engage directly with evolving AI technologies. These experiences may include prompt-design exercises, comparative analyses of outputs generated by different AI platforms, and evaluation of factual accuracy, bias, and hallucinations. Such activities encourage students to become critical consumers of AI-generated content rather than passive recipients.<\/p>\n

Discipline-Specific AI Case Analysis<\/h3>\n

Students can investigate contemporary organizational, business, healthcare, engineering, and public-sector applications of AI. By examining current use cases, emerging capabilities, regulatory developments, and ethical considerations, students gain a more nuanced understanding of AI’s role across industries. These analyses may be conducted individually or collaboratively and shared through written submissions, presentations, or facilitated class discussions. Faculty should encourage students to evaluate not only the opportunities presented by AI but also its limitations, uncertainties, and potential risks within their discipline.<\/p>\n

Developing AI Evaluation Literacy<\/h3>\n

Given the limitations of current AI-detection technologies and the prevalence of false positives, students should be educated on the strengths and weaknesses of AI-evaluation tools. Rather than treating AI detectors as definitive sources of truth, students should learn to identify writing patterns, assess transparency of AI use, and evaluate the quality and originality of their own work. Developing this literacy encourages responsible use while supporting academic integrity.<\/p>\n

AI Transparency and Reflective Practice<\/h3>\n

Major assignments can incorporate AI disclosure statements\/declarations that require students to document the nature and extent of AI use. These reflections may include the tools utilized, specific tasks performed with AI assistance, perceived benefits, limitations encountered, and lessons learned throughout the process. Such practices promote transparency while encouraging metacognitive reflection about learning. They also help faculty distinguish appropriate AI-assisted learning from excessive dependence on AI-generated work.<\/p>\n

Independent Inquiry and Knowledge Sharing<\/h3>\n

Students may investigate emerging AI applications, compare multiple AI tools, and evaluate how different approaches influence the quality, reliability, and efficiency of their work. Sharing findings with peers creates a collaborative learning environment while helping students recognize the evolving relationship between AI technologies and their future careers.<\/p>\n

Faculty and Industry Partnerships<\/h3>\n

Inviting faculty members from multiple disciplines, alongside industry practitioners, can expose students to diverse perspectives on AI implementation. Guest speakers can help connect technical developments, ethical considerations, and workplace expectations to students’ particular fields of study. Industry engagement helps students understand how AI is currently being used in professional practice and reinforces that successful AI adoption requires domain expertise, engineering judgment, ethical decision-making, and effective communication.<\/p>\n

Alumni Engagement<\/h3>\n

Recent graduates offer a particularly valuable perspective because they can speak directly to the transition from college to the workforce. Alumni can provide practical insights regarding AI adoption in professional settings, employer expectations, and the competencies students should cultivate to remain competitive and effective in their careers.<\/p>\n

Leveraging Institutional Support Resources<\/h3>\n

Institutions can further strengthen AI literacy by encouraging students to engage with existing campus resources such as writing centers, libraries, tutoring services, and dedicated AI learning labs. These resources can help students develop both technical proficiency and critical evaluation skills.<\/p>\n

In Conclusion<\/h3>\n

Collectively, these instructional approaches emphasize critical evaluation, reflective practice, and human judgment. They encourage students to engage with AI as a tool that augments learning and professional practice rather than one that replaces critical thinking, disciplinary expertise, or human judgment. In doing so, higher education can promote intentional forms of AI use while discouraging excessive dependence on AI for cognitive tasks that are central to student development. Developing AI literacy in this way is also an important component of workforce readiness, as students must learn not only how to use AI tools effectively, but also when their use is appropriate, how to critically evaluate AI-generated information, and when professional judgment and human expertise must take precedence.<\/p>\n

Ultimately, these practices can help narrow the gap between employer expectations and perceived graduate preparedness. By intentionally integrating AI literacy into teaching and learning, colleges and universities can better prepare students for an increasingly AI-enabled workplace. In this context, AI literacy extends beyond technical proficiency to include the critical thinking, ethical decision-making, and professional judgment needed to evaluate AI-generated information and make responsible decisions within professional settings. Positioning AI literacy as a component of workforce readiness can help ensure that graduates are prepared not simply to use emerging technologies, but to exercise the disciplinary knowledge, intellectual independence, and professional judgment necessary to determine how and when those technologies should be used.<\/p>\n

5. AI Literacy Outcomes for Students<\/h2>\n

AI literacy should be understood as a multidimensional educational outcome, not simply the ability to access an AI tool or write an effective prompt. It includes sufficient knowledge of how AI systems use data and generate outputs; the practical ability to select, operate, and integrate appropriate tools; the capacity to evaluate their performance and limitations; and an understanding of their ethical, legal, and societal implications. Hackl et al. (2026) organize these capacities into seven interconnected dimensions: technical knowledge, application proficiency, critical thinking, ethical reasoning, social-impact understanding, integration skills, and legal and regulatory knowledge. Similarly, Kang and Park\u2019s (2026) validated scale for higher education students identifies four related dimensions: fundamental knowledge, practical application, performance evaluation, and impact assessment. Although Zhou et al. (2025) distinguish AI literacy, or what individuals know, from AI competency, or how effectively and reflectively they apply that knowledge, higher education should cultivate both.<\/p>\n

An AI-literate graduate should therefore be able to explain, at an appropriate level, how AI systems are trained and why their outputs can be incomplete, inaccurate, biased, or fabricated. Students should be able to identify when AI is suitable for a task, formulate and refine productive queries, and integrate its outputs with disciplinary knowledge without surrendering human judgment. They should know how to verify claims against credible evidence, compare alternative outputs, recognize uncertainty and embedded assumptions, and determine whether AI has strengthened or displaced their own reasoning. AI literacy also requires students to protect personal and confidential information, respect intellectual property, document and disclose their use of AI, and accept responsibility for the resulting work. Finally, students should be prepared to evaluate AI\u2019s broader consequences for fairness, access, employment, human relationships, democratic life, and the environment. Because the technology and its governing rules will continue to change, students must also learn to assess their own capabilities, recognize overreliance, and update their knowledge over time (Zhou et al., 2025). These outcomes should be taught and assessed throughout the curriculum rather than relegated to academic-integrity policies or optional technical training.<\/p>\n

6. Conclusion and Resources<\/h2>\n

Responsible AI integration is fundamentally a pedagogical challenge that higher education institutions must address proactively and intentionally. As AI becomes increasingly embedded across professional sectors, employers will expect college graduates to possess both AI proficiency and the critical-thinking skills necessary to evaluate, apply, and question AI-generated information. Therefore, the development of AI literacy should not occur at the expense of the cognitive skills and disciplinary knowledge that students need to exercise independent and professional judgment.<\/p>\n

Higher education institutions have an important role in creating learning environments in which students develop AI proficiency alongside critical thinking, ethical decision-making, and responsible technology use. Rather than focusing solely on whether students should or should not use AI, educators should consider how AI can be intentionally integrated into teaching and learning in ways that support specific learning objectives while preserving opportunities for students to engage in independent thinking, problem-solving, and knowledge development.<\/p>\n

Collaboration among faculty, students, and institutions is essential to this process. Establishing clear expectations for responsible and ethical AI use, developing students\u2019 technical and critical AI literacy, and increasing awareness of how AI is being used across professional sectors can help demystify emerging technologies while preparing students to use them effectively and responsibly. Ultimately, higher education must prepare graduates not simply to use AI, but to determine when and how it should be used, critically evaluate its outputs and limitations, and exercise the human judgment necessary to make ethical and informed decisions in an increasingly AI-mediated workforce.<\/p>\n

Resources for Responsible AI Integration<\/h3>\n

Several resources can support institutions and faculty as they navigate the responsible integration of AI into teaching and learning. The EDUCAUSE AI Literacy in Teaching and Learning: A Durable Framework for Higher Education<\/em> provides a comprehensive framework for developing AI literacy among students, faculty, and staff. The framework emphasizes understanding AI fundamentals, critically evaluating AI tools and outputs, recognizing limitations and potential biases, and using AI effectively and ethically in academic and professional contexts (Kassorla et al., 2024).<\/p>\n

EDUCAUSE also maintains a Teaching with Artificial Intelligence<\/em> resource collection that brings together practical materials related to classroom AI use, academic integrity, course design, AI literacy, and policy development (EDUCAUSE, n.d.). These resources can assist faculty and institutional leaders in translating broader principles of responsible AI use into teaching and learning practices.<\/p>\n

Additionally, the University of Georgia Center for Teaching and Learning provides practical guidance for developing generative AI policies at the course and assignment levels. Vanderveen and Poproski (2026.) emphasize the importance of clearly communicating AI expectations and their rationale while recognizing that different learning goals may warrant different levels of AI use within the same course. This approach provides faculty with a practical model for aligning AI expectations with learning objectives rather than relying on a single universal policy for all learning activities.<\/p>\n

Together, these resources offer faculty and institutions practical starting points for developing AI literacy, establishing transparent policies, aligning AI use with learning objectives, and maintaining human judgment and critical thinking at the center of teaching and learning.<\/p>\n

7. AI Statement<\/h2>\n

Section 1: AI Statement: Researched and drafted by a human being without AI assistance; ChatGPT assisted with editing and revision.<\/p>\n

Section 2: AI Statement: Researched and drafted by a human being without AI assistance; ChatGPT assisted with editing and revision.<\/p>\n

Section 3: Researched and drafted by a human being without AI assistance; ChatGPT assisted with editing and revision.<\/p>\n

Section 4: AI Statement: Researched and drafted by a human being without AI assistance; CoPilot assisted with editing and revision.<\/p>\n

Section 5: AI Statement: Researched and drafted by a human being without AI assistance; ChatGPT assisted with editing and revision.<\/p>\n

References<\/h2>\n

Aguilar, S. J., Nye, B., Swartout, W. R., Macias, A., Xing, Y., & Xiu, R. (2025, August). Fostering critical thinking in the age of AI<\/em>. <\/a>https:\/\/doi.org\/10.35542\/osf.io\/wr6n3_v2<\/a><\/p>\n

Barus, O. P., Hidayanto, A. N., Handri, E. Y., Sensuse, D. I., & Yaiprasert, C. (2025). Shaping generative AI governance in higher education: Insights from student perception. International Journal of Educational Research Open, 8<\/em>, Article 100452. https:\/\/doi.org\/10.1016\/j.ijedro.2025.100452<\/a><\/p>\n

Bassner, P., Lenk-Ostendorf, B., Beinstingel, R., Wasner, T., & Krusche, S. (2026). Less stress, better scores, same learning: The dissociation of performance and learning in AI-supported programming education. Computers and Education: Artificial Intelligence, 10<\/em>, Article 100537. <\/a>https:\/\/doi.org\/10.1016\/j.caeai.2025.100537<\/a><\/p>\n

Bastani, H., Bastani, O., Sungu, A., Ge, H., Kabakci, O., & Mariman, R. (2025). Generative AI without guardrails can harm learning: Evidence from high school mathematics. Proceedings of the National Academy of Sciences, 122<\/em>(26), Article e2422633122. <\/a>https:\/\/doi.org\/10.1073\/pnas.2422633122<\/a><\/p>\n

Bauer, E., Greiff, S., Graesser, A. C., Scheiter, K., & Sailer, M. (2025). Looking beyond the hype: Understanding the effects of AI on learning. Educational Psychology Review, 37<\/em>, Article 45. <\/a>https:\/\/doi.org\/10.1007\/s10648-025-10020-8<\/a><\/p>\n

Chirkov, I., Smirnov, I., & Kizilcec, R. F. (2026). Generative AI use and misuse call for assessment reform in higher education. Science, 392<\/em>(6800), 818\u2013820. <\/a>https:\/\/doi.org\/10.1126\/science.aec5115<\/a><\/p>\n

Contractor, Z., & Reyes, G. (2026). Experimental evidence on the learning impact of generative AI<\/em> (IZA Discussion Paper No. 18792). IZA Institute of Labor Economics. <\/a>https:\/\/www.iza.org\/publications\/dp\/18792<\/a><\/p>\n

Cooper, M. M. (2026). A preliminary set of principles to support learning in the context of generative AI. Journal of Chemical Education, 103<\/em>, 36\u201342. <\/a>https:\/\/doi.org\/10.1021\/acs.jchemed.5c01413<\/a><\/p>\n

EDUCAUSE. (n.d.). Teaching with artificial intelligence<\/em>. Teaching with Artificial Intelligence<\/a><\/p>\n

Elosh, E., Kumar, S., & Sukhoy, A. (2025). Integrating AI at CSU: Exploring the current state and future opportunities with our faculty, students and alumni.<\/em><\/p>\n

Farhat, Z. (2026). When AI thinks for us: Unveiling the cognitive and social toll of ChatGPT over-reliance. Applied Cognitive Psychology, 40<\/em>, Article e70205. <\/a>https:\/\/doi.org\/10.1002\/acp.70205<\/a><\/p>\n

Hackl, V., M\u00fcller, A. E., & Sailer, M. (2026). The AI literacy heptagon: A structured approach to AI literacy in higher education. Computers and Education: Artificial Intelligence, 10<\/em>, Article 100540. <\/a>https:\/\/doi.org\/10.1016\/j.caeai.2026.100540<\/a><\/p>\n

Hon, K. (2026). Generative AI in higher education: A systematic review of its effects on learning outcomes and academic performance. Journal of Educational Technology Systems, 54<\/em>(3), 537\u2013560. <\/a>https:\/\/doi.org\/10.1177\/00472395251400089<\/a><\/p>\n

Jose, B., Cherian, J., Verghis, A. M., Varghise, S. M., S, M., & Joseph, S. (2025). The cognitive paradox of AI in education: Between enhancement and erosion. Frontiers in Psychology, 16<\/em>, Article 1550621. <\/a>https:\/\/doi.org\/10.3389\/fpsyg.2025.1550621<\/a><\/p>\n

Kang, K. Y., & Park, J.-H. (2026). Development and validation of a multidimensional AI literacy scale for higher education students: A mixed-method study. Journal of Librarianship and Information Science<\/em>. Advance online publication. <\/a>https:\/\/doi.org\/10.1177\/09610006261425235<\/a><\/p>\n

Kassorla, M., Georgieva, M., & Papini, A. (2024). AI literacy in teaching and learning: A durable framework for higher education<\/em>. EDUCAUSE. AI Literacy in Teaching and Learning<\/a><\/p>\n

Lodge, J. M., & Loble, L. (2026). Artificial intelligence, cognitive offloading and implications for education<\/em>. University of Technology Sydney. <\/a>https:\/\/doi.org\/10.71741\/4pyxmbnjaq.31302475<\/a><\/p>\n

Pallant, J. L., Blijlevens, J., Campbell, A., & Jopp, R. (2026). Mastering knowledge: The impact of generative AI on student learning outcomes. Studies in Higher Education, 51<\/em>(4), 714\u2013735. <\/a>https:\/\/doi.org\/10.1080\/03075079.2025.2487570<\/a><\/p>\n

Slade, J. J. (2025). How to help your students use AI without losing the learning<\/em>. American Psychological Association. <\/a>https:\/\/www.apa.org\/ed\/precollege\/psychology-teacher-network\/introductory-psychology\/learning-artificial-intelligence<\/a><\/p>\n

Vanderveen, T., & Poproski, R. (2026). Generative AI policies for teaching & learning<\/em>. University of Georgia Center for Teaching and Learning. Generative AI Policies for Teaching & Learning<\/a><\/p>\n

Wang, H., Dang, A., Wu, Z., & Mac, S. (2024). Generative AI in higher education: Seeing ChatGPT through universities’ policies, resources, and guidelines. Computers and Education: Artificial Intelligence, 7<\/em>, Article 100326. https:\/\/doi.org\/10.1016\/j.caeai.2024.100326<\/a><\/p>\n

Zhou, X., Li, Y., Chai, C. S., & Chiu, T. K. F. (2025). Defining, enhancing, and assessing artificial intelligence literacy and competency in K\u201312 education from a systematic review. Interactive Learning Environments, 33<\/em>(10), 5766\u20135788. <\/a>https:\/\/doi.org\/10.1080\/10494820.2025.2487538<\/a><\/p>\n

 <\/p>\n","protected":false},"parent":0,"menu_order":6,"template":"","meta":{"pb_part_invisible":false},"contributor":[],"license":[],"class_list":["post-112","part","type-part","status-publish","hentry"],"_links":{"self":[{"href":"https:\/\/pressbooks.ulib.csuohio.edu\/usingaiinacademics\/wp-json\/pressbooks\/v2\/parts\/112","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/pressbooks.ulib.csuohio.edu\/usingaiinacademics\/wp-json\/pressbooks\/v2\/parts"}],"about":[{"href":"https:\/\/pressbooks.ulib.csuohio.edu\/usingaiinacademics\/wp-json\/wp\/v2\/types\/part"}],"version-history":[{"count":6,"href":"https:\/\/pressbooks.ulib.csuohio.edu\/usingaiinacademics\/wp-json\/pressbooks\/v2\/parts\/112\/revisions"}],"predecessor-version":[{"id":276,"href":"https:\/\/pressbooks.ulib.csuohio.edu\/usingaiinacademics\/wp-json\/pressbooks\/v2\/parts\/112\/revisions\/276"}],"wp:attachment":[{"href":"https:\/\/pressbooks.ulib.csuohio.edu\/usingaiinacademics\/wp-json\/wp\/v2\/media?parent=112"}],"wp:term":[{"taxonomy":"contributor","embeddable":true,"href":"https:\/\/pressbooks.ulib.csuohio.edu\/usingaiinacademics\/wp-json\/wp\/v2\/contributor?post=112"},{"taxonomy":"license","embeddable":true,"href":"https:\/\/pressbooks.ulib.csuohio.edu\/usingaiinacademics\/wp-json\/wp\/v2\/license?post=112"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}