# How university professors can use artificial intelligence > Discover how university professors can use artificial intelligence to plan classes, create activities, support students and improve their teaching practice. - Site: Genialoh (https://genialoh.org) - Language: en - Category: Teaching - Reading time: 14 min - HTML version: https://genialoh.org/#/en/blog/artificial-intelligence-for-university-professors --- Artificial intelligence can help university professors prepare classes, create materials, design activities, support students and analyze different ways to explain a concept. However, using AI in teaching does not mean automatically delegating the professor's work or accepting any generated answer as correct. Value appears when the teacher uses the technology with a pedagogical purpose, reviews the results, protects institutional information and keeps responsibility over academic decisions. For a university, the challenge is not just that some professors learn to use AI tools. The challenge is to create common criteria so that adoption is useful, responsible and coherent with the educational model. In this guide we present concrete uses of artificial intelligence for university professors, their main limits and a path to integrate them inside an institutional strategy. ## What can artificial intelligence bring to a university professor? AI can work as an assistant that helps to explore alternatives, organize information, prepare drafts and generate initial materials that will then be reviewed by the professor. It can add value in five areas: - Class planning. - Materials creation. - Design of activities and assessments. - Student support. - Development and improvement of teaching practice. The tool does not replace disciplinary knowledge, pedagogical judgment or the professor-student relationship. > **The professor remains responsible for** > > - Defining learning objectives. > - Selecting content. > - Verifying information. > - Adapting materials. > - Setting the rules of use. > - Evaluating learning. > - Protecting data. > - Making final decisions. ## 1. Prepare the structure of a class A professor can use AI to develop a first session structure. They can provide topic, student level, class duration, learning objective, prior knowledge, activity type and available resources. From this context, AI can propose a sequence with introduction, explanation, practice, discussion and closing. > **Example prompt** > > Propose a structure for a 90-minute class on cost analysis for second-semester students. Include an opening activity, an example, teamwork and a final reflection. The generated proposal should be treated as a draft. The professor needs to review depth, correctness of concepts, realistic timing, alignment of activities with the objective, relevance of examples and compatibility with their methodology. ## 2. Explore different ways to explain a concept The same topic can be clear for some students and confusing for others. AI can help the professor explore alternative explanations: - A formal definition. - A simple explanation. - An analogy. - A professional example. - An everyday case. - A visual explanation. - A comparison with a known concept. - A step-by-step sequence. A statistics teacher, for example, can ask for different ways to explain the difference between correlation and causation, then choose, correct and adapt the versions that best fit the group. AI does not automatically know the specific difficulties of the students. The professor must provide level, context and observed errors. ## 3. Create examples and counterexamples Examples help apply abstract concepts. Counterexamples help recognize limits, errors and misinterpretations. A professor can use AI to generate correct cases, incorrect cases, ambiguous situations, examples from different sectors, scenarios with different difficulty levels and comparisons between good and bad decisions. In a law class, they could ask for examples where an argument seems valid but contains an incorrect premise. In engineering, for problems with frequent errors so students identify them. In communication, for messages that are appropriate and inappropriate for different audiences. All examples must be reviewed before being presented. The tool can invent data, oversimplify or produce situations that do not properly reflect the discipline. ## 4. Design discussion questions AI can help create questions that promote analysis, reflection and participation: comprehension, application, comparison, evaluation, argumentation, decision making, ethical reflection and professional practice. > **Example prompt** > > Generate ten questions to discuss the ethical implications of using AI in hiring processes. Order them from least to most complex. The professor should select questions aligned with the session objectives and adapt overly generic ones to the local context, student experience or real professional cases. ## 5. Prepare learning activities AI can help generate ideas for individual or collaborative activities: case analysis, debates, simulations, concept maps, problem solving, comparing alternatives, short projects, error-identification exercises, research activities and reflections on professional practice. The professor must state which competence they want to develop. Asking for “an activity about leadership” is not the same as: design a 40-minute activity in which management students compare three leadership styles, apply criteria to a case and defend a recommendation. The clearer the pedagogical purpose, the more useful the generated draft. ## 6. Adapt an activity to different levels Within the same group there can be diversity in prior knowledge, experience and learning pace. AI can help prepare different versions of an activity. **Variants of a single activity** | Variant | Focus | | --- | --- | | Introductory | Base concepts and guided examples. | | Intermediate | Application with partial support. | | Advanced | Challenge with more autonomy and complexity. | | More support | Detailed instructions and additional scaffolding. | | More autonomy | Open cases and student decisions. | This does not mean creating a completely different plan for each student. It can consist of offering different entry points or levels of support around the same learning objective. The professor should avoid adaptations that unjustifiably lower academic expectations. The difference should lie in support or gradual complexity, not in removing fundamental learning. ## 7. Generate practice questions AI can prepare additional questions for students to practice before an assessment: multiple choice, open questions, application exercises, problems, short cases, true or false, matching activities and reflection prompts. The professor must carefully review accuracy, difficulty level, clarity of instructions, quality of options, whether more than one answer is possible, alignment with the syllabus and correspondence with what was taught. Generated questions should not be added directly to a high-stakes assessment without professional review. ## 8. Create initial rubrics A rubric helps communicate the criteria used to evaluate a product or performance. AI can generate a first draft based on learning objective, activity type, dimensions to evaluate, number of levels, characteristics of each level and weighting. > **Example prompt** > > Create a draft four-level rubric to evaluate a business project presentation. Include problem analysis, use of evidence, feasibility, communication and response to questions. The professor must ensure that criteria are observable, differentiable and relevant, and remove vague formulations like “excellent”, “adequate” or “poor” when they lack concrete descriptions. ## 9. Review clarity of instructions and materials A professor can use AI to identify possible problems in their own instructions, guides or materials: ambiguities, missing steps, overly technical language, repetitions, contradictions, difficulty level, unclear delivery criteria and unexplained assumptions. > **Example prompt** > > Review these instructions as if you were a first-semester student. Identify which parts may be confusing and propose questions the student might ask. The professor does not have to accept the corrections automatically. The review acts as a second reading to detect possible sources of confusion. ## 10. Prepare support materials AI can help prepare drafts of summaries, glossaries, study guides, FAQs, concept lists, review cards, timelines, comparisons, topic introductions and materials for makeup sessions. These resources must be based on verified content aligned with the syllabus. When the university has an institutional platform contextualized with its curricula, programs and authorized materials, these supports can keep institutional guidelines present. ## 11. Anticipate frequent errors The professor can ask AI to propose common errors associated with a concept, procedure or activity: misreading a chart, confusing legal concepts, failing at a mathematical procedure, poorly formulating a hypothesis, bad practices in an interview, citation errors or incorrect assumptions in a financial analysis. The teacher must check that those errors are real and relevant for their group. A particularly useful way to apply this function is to provide anonymized examples of errors already observed, without student names or personal information. ## 12. Prepare formative feedback AI can help draft a first version of feedback, especially when the professor wants to organize comments around common criteria. It can help structure what has been achieved, what needs improvement, the evidence supporting the comment, a reflection question, a next step and a resource to continue. However, automatic feedback should not be sent without review. The professor must verify that the comment truly corresponds to the work, no non-existent errors are attributed, the tone is appropriate, the recommendation is applicable, the student's context is respected and no confidential information is entered into an unauthorized tool. Educational feedback also has a human function. It recognizes effort, decisions and particular circumstances that a tool may not understand. ## 13. Design activities on the critical use of AI AI is not only used to create materials. It can also become a learning object. Professors can design activities where students: - Compare two generated answers. - Identify errors. - Verify claims. - Analyze references. - Detect bias. - Improve an explanation. - Document changes. - Justify which suggestions they accepted or rejected. - Compare an AI answer with academic sources. - Reflect on the tool's limits. This approach develops judgment and AI literacy. Instead of asking students to trust or distrust in general, it prepares them to evaluate results by evidence, purpose and context. ## 14. Redesign activities that can be easily delegated Some traditional tasks can be almost fully completed with generative tools. The institutional response is not limited to trying to detect use. The professor can review which learning evidence they want to observe. An activity can be strengthened through application to a local case, oral explanation, intermediate deliveries, a decision log, comparison of sources, personal reflection, student-collected data, defense of a recommendation, error review and declaration of AI use. Instead of asking only for a general essay, one can request: initial formulation of the argument, comparison with a generated answer, source verification, text revision and explanation of the decisions made. This way, AI is part of the process without hiding learning evidence. ## 15. Prepare a class with an institutional platform An institutional AI platform organizes the experience around the university, not just around each user's individual preferences. The professor can work with tutors configured by program, context based on the curriculum, academic programs, institutional methodology, values, guidelines, authorized materials and institution-defined use cases. This helps professors and students share common references and allows training, support and monitoring to develop as part of an institutional project. ## What information a professor should not enter Professors must avoid sharing personal, sensitive or confidential information in unauthorized tools. > **They should not enter** > > - Full student names. > - Student IDs. > - Records. > - Individual grades. > - Medical information. > - Financial information. > - Identifiable personal situations. > - Family data. > - Passwords. > - Confidential documents. > - Materials without authorization to share. When using examples of student work or errors, they must be properly anonymized. Specific obligations depend on jurisdiction, university policies and the tool used, so the professor must know and follow institutional guidelines. ## Seven principles for using AI in university teaching | Principle | Key idea | | --- | --- | | 1. Start with the learning objective | The question is not where to use AI, but what students need to learn and how it can help. | | 2. Verify before using | AI can produce incorrect, incomplete or invented information. All academic content must be checked. | | 3. Keep human responsibility | Decisions about content, activities, evaluation and feedback belong to the professor. | | 4. Protect information | Do not enter personal, sensitive or institutional data in unauthorized tools. | | 5. Communicate the rules | Students must know when AI is forbidden, allowed or required. | | 6. Preserve learning evidence | Activities must show the student's reasoning, decisions and understanding. | | 7. Document and improve | Record use cases, difficulties, errors and adjustments for institutional learning. | ## How to set rules for students Each activity must clearly state which use of AI is allowed. A practical classification can be: | Level | Description | | --- | --- | | AI not allowed | Student completes the activity without generative tools. Useful for foundational knowledge or individual performance. | | AI allowed with restrictions | Can be used in specific phases (generating questions, checking clarity, exploring examples). The professor states what remains the student's sole responsibility. | | AI allowed with disclosure | The student can use it but explains tool, purpose, prompts, changes made and verification. | | AI required | The activity deliberately uses the tool to analyze its outputs or develop professional competencies. | The rule must be communicated before starting, not after the delivery. ## Training professors need AI training for teachers should include foundations of generative AI, capabilities and limits, information verification, prompt design, pedagogical use cases, academic integrity, activity redesign, data protection, use of institutional platforms and evaluation and monitoring. A single conference can spark interest but is usually not enough to transform practice. It is better to combine a general session, a hands-on workshop, discipline-based work, application in a course, follow-up, experience sharing and review of examples. ## How to evaluate whether AI is helping the professor The university can observe trained professors, professors using the platform, identified use cases, materials created or adapted, redesigned activities, perceived usefulness, time spent on certain tasks, difficulties, support requests, incidents, training needs and improvement recommendations. The number of interactions should not be automatically read as academic impact. Evaluation must combine quantitative information with interviews, surveys, material review and documented experiences. ## Frequent mistakes when using AI in teaching | Mistake | Why to avoid it | | --- | --- | | Copying the output without reviewing | A well-written answer can contain errors or non-existent references. | | Using AI without a pedagogical goal | The tool can add complexity without improving learning. | | Entering student information | Convenience does not justify sharing personal data in unauthorized systems. | | Generating assessments without checking them | Questions can be ambiguous, incorrect or have several valid answers. | | Fully automating feedback | Students may receive generic comments disconnected from their real work. | | Requiring uniform use across professors | Each discipline needs different use cases and limits. | | Presenting it as infallible | Students must know its limits and learn to verify. | | Ignoring adoption workload | Implementation requires training, practice, support and time to adjust. | ## Example four-week teaching plan | Week | Focus | Activities | | --- | --- | --- | | 1 | Exploration | Get to know the platform, test explanations, identify errors, review guidelines, select a use case. | | 2 | Design | Adapt an activity, define student rules, prepare materials, set evaluation criteria. | | 3 | Application | Present the activity, support students, resolve questions, document incidents. | | 4 | Evaluation | Review results, gather feedback, identify adjustments, share learnings with coordination. | ## Frequently asked questions ### Can AI replace a university professor? It should not be framed as a replacement. Teaching requires disciplinary judgment, pedagogical design, evaluation, support and human responsibility. ### Can AI be used for grading? It can help organize criteria or draft comments, but grading decisions must keep human supervision and responsibility. Privacy, equity and institutional policy also apply. ### Can professors share student work? They should not share identifiable work in unauthorized tools. Any example must be anonymized and used according to institutional policies. ### Can AI generate correct materials automatically? No. It can generate useful drafts, but also errors, oversimplifications, bias or non-existent references. All material must be reviewed. ### Which professors should start? Those with clear use cases, willingness to attend training and capacity to document the experience. ### Do all professors need to know how to code? No. Most educational uses do not require programming. What matters is understanding objectives, limits, verification and institutional rules. ### How can a university avoid every professor using different tools? It can define an institutional strategy, set authorized tools, provide its own platform, train teachers and keep a process to evaluate new solutions. ### Does Genialoh offer support for professors? Among Genialoh's confirmed capabilities are AI support for professors, contextualization with the curriculum, teacher training, implementation, support and adoption monitoring. ## From isolated tools to an institutional teaching strategy AI can help a professor prepare classes, explore explanations, design activities and support students. But its real institutional value appears when those uses connect with the university's programs, methodology, values and guidelines. > **With Genialoh you can create an AI platform with** > > - The institution's name, logo and colors. > - Context based on curricula and programs. > - Tutors configured by program. > - AI support for professors. > - Personalized learning. > - Institutional reports and analytics. > - Teacher training, implementation, support and adoption monitoring. Book a first 30-minute meeting to get to know Genialoh. If your institution decides to move forward, we will select two programs and prepare a functional demo with its identity and academic context in less than 14 days from receiving the complete documentation. After the presentation, up to five authorized people can explore the demo for 30 calendar days.