# How to train university professors in artificial intelligence: a leveled training plan

> Design a training plan on artificial intelligence for university professors with levels, practical workshops, assessment, coaching and adoption metrics.

- Site: Genialoh (https://genialoh.org)
- Language: en
- Category: Faculty development
- Reading time: 24 min
- HTML version: https://genialoh.org/#/en/blog/train-university-professors-artificial-intelligence

---
Training university professors in artificial intelligence is not just teaching them how to write prompts for a generative tool.

An institutional training plan must prepare teachers to understand how the technology works, recognize its limits, protect information, design activities, update assessment and use AI without delegating their academic responsibility.

It must also consider that teachers do not start at the same point. Within a single university you may find:

- Teachers who have never used a generative tool.
- Professors who use it occasionally.
- Frequent users who do not yet apply common pedagogical criteria.
- Teachers who already design activities with AI.
- Coordinators able to coach other teachers.
- Specialists interested in building tutors and use cases by program.

A single two-hour talk will rarely be enough. The university needs a leveled plan combining diagnostic, fundamentals, practice, discipline-based application, activity design, academic integrity, privacy, assessment, use of institutional platforms, coaching, measurement and continuous improvement.

Genialoh includes AI support for teachers, faculty training, curricular personalization, implementation and adoption follow-up as components of its institutional proposal.

## Why a university needs to train its professors in AI

Students already use AI to look up explanations, solve doubts, generate ideas, write texts, review code, prepare presentations, translate, summarize, study and practice.

Meanwhile, teachers face questions such as:

- Which uses should I allow?
- How do I know whether the student learned?
- Do I need to redesign my activities?
- Can I use AI to prepare a class?
- What information should I not share?
- How do I verify an answer?
- How do I assess AI-assisted work?
- Can I use it for feedback?
- What happens if it generates false information?
- Which tool is authorized?
- How does it relate to the curriculum?
- Which institutional rules should I apply?

Without common training, each teacher responds differently. This can produce contradictory rules, hard-to-enforce bans, unsupervised use, incorrect materials, privacy risks, assessments that no longer show learning, faculty resistance, scattered tools, lack of adoption evidence and inequality between programs.

Training should not force every teacher to use AI in the same way. It should provide a common baseline so each discipline decides where it adds value, where it needs limits and where it should not be used.

## What an institutional training plan should achieve

At the end of a complete route, teachers should be able to:

- Explain what generative AI is and its limits.
- Identify use cases relevant to their course.
- Write prompts with context and criteria.
- Verify generated content.
- Protect data and institutional documents.
- Define usage rules for students.
- Design activities that preserve evidence of learning.
- Adapt assessment criteria.
- Use AI as support without delegating academic decisions.
- Use an institutional platform.
- Document use cases, problems and learnings.
- Know when not to use AI.
- Recognize situations that require human intervention.
- Participate in improving the institutional strategy.
- Coach other teachers when appropriate.

## Training must start with a diagnostic

Before designing workshops, the university needs to understand the starting point. A short diagnostic can explore usage frequency, tools used, current use cases, confidence level, main doubts, risk perception, knowledge of policies, experience in activity design, needs by discipline, availability to participate and interest in coaching others.

### Recommended diagnostic survey

- Have you used generative AI tools?
- How often?
- For which tasks?
- Have you used them in teaching activities?
- Have you allowed students to use them?
- How do you communicate the rules?
- Have you modified any assessment because of AI?
- Do you know what data must not be shared?
- Do you feel able to verify a generated answer?
- What worries you most?
- Which use case would you like to learn?
- Which training modality do you prefer?
- How much time can you dedicate?
- Would you be interested in serving as a mentor?

> **Avoid rigid labels**
> 
> The diagnostic should not simplistically classify teachers as «resistant» or «innovators». One person may use AI frequently but ignore privacy risks. Another may use it little but deeply understand pedagogical design.

## Four-level training model

| Level | Focus | Final product |
| --- | --- | --- |
| 1. Literacy | Understand, experiment and protect information | Verified practice |
| 2. Application | Prepare materials and teaching tasks | Applied resource |
| 3. Pedagogical design | Redesign activities and assessment | Activity with rubric |
| 4. Leadership | Coach, document and scale | Institutional project |

## Level 1. Literacy and responsible use

### Goal

That teachers understand what generative AI can and cannot do, know the institutional rules and safely carry out their first practices.

### Suggested duration

Four to six hours, over two or three sessions.

### Module 1. What generative AI is

In accessible language, it must explain what kind of responses AI produces, how it uses patterns, why it can respond fluently, why it can be wrong, what hallucinations are, how context modifies results, differences between searching and generating text, and what responsibility the user retains.

The goal is not to train technical specialists. It is to avoid two extremes: believing AI is infallible, or considering it has no academic use.

### Module 2. Capabilities and limits

Teachers should try tasks such as explaining a concept, proposing examples, creating questions, reviewing instructions, organizing ideas and generating a draft. Then they should identify errors, omissions, simplifications, assumptions, non-existent references, level mismatches and verification needs.

> **Recommended activity**
> 
> Give all teachers a generated response containing an incorrect fact, a non-existent reference, a simplification and a poorly supported conclusion. The group must locate the problems and explain how they would verify them.

### Module 3. Information protection

Teachers must know what data not to enter into unauthorized tools: student names, IDs, records, individual grades, medical information, financial information, family data, identifiable personal situations, passwords, confidential exams, question banks, unpublished research, strategic institutional documents and unauthorized materials.

> **Case example**
> 
> A teacher wants help drafting feedback and uploads a full student paper with name and ID. The group must discuss what information is unnecessary, how it could be anonymized, whether the tool is authorized, what safer alternative exists and who retains responsibility for the comment.

### Module 4. First prompts

Teachers can learn a simple structure: role, context, goal, task, restrictions, format and criteria.

> **Example**
> 
> Act as an instructional design assistant. The activity is for second-semester business students. The goal is to distinguish fixed and variable costs. Propose three short cases without providing the answers. Present each case with context, data and a final question.

### Evidence to complete the level

- A contextualized prompt.
- The generated output.
- Three problems identified.
- A version corrected by the teacher.
- A brief reflection on appropriate use.

## Level 2. Teaching application

### Goal

That teachers use AI to prepare materials and support teaching tasks without delegating academic decisions.

### Suggested duration

Six to eight hours, combining workshops and individual work.

### Module 1. Class planning

Teachers can practice using AI to structure sessions, formulate objectives, organize content, propose activities, design closings, prepare questions and estimate timing. Each teacher selects a real class and generates an initial draft, then reviews program alignment, sequence, level, timing, participation, formative assessment and resources. The final version must clearly show what the teacher changed.

### Module 2. Explanations, examples and practice

AI can support the initial creation of alternative explanations, analogies, examples, counterexamples, cases, exercises, practice questions and frequent errors. The teacher must verify accuracy, relevance, language, complexity, disciplinary fit and possible biases.

### Module 3. Reviewing materials

Participants can use AI as a second read for instructions, guides, rubrics, presentations, activities, FAQs and support materials. The tool can flag ambiguities, but the teacher decides which changes are appropriate.

### Module 4. Formative feedback

AI can help prepare a draft organized by strengths, areas to improve, evidence, reflection question and next step. It should not be sent automatically. The teacher must anonymize any example, review correspondence with the work, adjust the tone, remove generic comments and retain final responsibility.

### Module 5. Discipline-based use cases

| Discipline | Use cases |
| --- | --- |
| Engineering and sciences | Procedure explanation, problem generation, code analysis, assumption review, edge cases, calculation verification |
| Business and administration | Case studies, alternative analysis, simulations, communication, risk evaluation, scenario preparation |
| Law and social sciences | Argument analysis, perspective comparison, concept review, fictional cases, bias identification, regulatory verification |
| Humanities and communication | Text analysis, style comparison, discussion questions, structure review, audiences, argumentation |
| Health | Simulated cases, communication with fictional patients, missing-information identification, decision analysis, controlled practice |

Health exercises should never be presented as real medical advice or replace supervised practice.

### Evidence to complete the level

- A material created with AI support.
- The prompt used.
- The initial output.
- Verifications performed.
- Teacher changes.
- An explanation of how it will be used.

## Level 3. Pedagogical design and assessment

### Goal

That teachers redesign activities and assessments to incorporate or limit AI according to the expected learning.

### Suggested duration

Eight to twelve hours, plus application in a course.

### Module 1. Define usage levels

| Level | Condition |
| --- | --- |
| AI not allowed | The student performs the activity without generative assistance |
| AI limited | The tool can only be used at specific stages |
| AI allowed with disclosure | It can be used, but must be documented |
| AI required | Using it critically is part of the learning |

The teacher must avoid ambiguous phrases like «Use AI responsibly». It is necessary to explain what that means in the activity.

### Module 2. Redesigning delegable activities

Teachers review tasks that a tool could complete almost entirely: general essays, summaries, definitions, descriptive reports, generic presentations, standard problems and context-free reflections. Then they incorporate local cases, student-collected data, partial deliveries, verification, sources, justified decisions, oral defense, reflection, application and process evidence.

> **Redesign example**
> 
> Original activity: Write an essay on AI effects on employment. Redesigned activity: select a local sector, formulate a question, generate two responses, identify verifiable claims, consult sources, detect errors, build an own argument, disclose use and defend the conclusions.

### Module 3. Assessing the process

The teacher can request drafts, logs, sources, main prompts, change table, verification, reflection, presentation and oral defense. It is not necessary to request every conversation. Evidence must be proportional.

### Module 4. Adapting rubrics

A rubric can include comprehension, analysis, verification, sources, application, justification, own production, transparency, reflection and defense.

### Module 5. Academic integrity

Teachers must learn to differentiate between authorized use, unauthorized use, lack of disclosure, misleading presentation, fabricated references, delegation of the central task, misuse of information and evidence manipulation. They should also understand that a detector score alone does not prove an infraction.

### Module 6. Use disclosures

> **Template**
> 
> I used [tool] for [purpose] during [stage]. I incorporated [elements]. I modified or rejected [explanation]. I verified the result through [sources or procedure]. I take responsibility for the final delivery.

### Evidence to complete the level

- A redesigned activity.
- AI usage level.
- Prompts.
- Required evidence.
- Rubric.
- Disclosure.
- Verification protocol.
- Reflection after applying it.

## Level 4. Leadership and institutional scaling

### Goal

Prepare teachers and coordinators to coach other faculty, create institutional resources and collaborate on improving the platform.

### Suggested duration

Twelve to twenty hours, plus an institutional project.

### Module 1. Facilitating communities of practice

Participants learn to run workshops, analyze cases, coach teachers, document good practices, avoid impositions, identify needs, escalate problems and share materials.

### Module 2. Creating institutional repositories

The university can develop a repository with reviewed prompts, activities, rubrics, cases, disclosures, FAQs, errors found, discipline examples, training materials and current rules. Each resource must include author, program, goal, date, tool, validation, limitations and version.

### Module 3. Designing tutors by career or grade

Teachers can collaborate in defining tutor purpose, level, programs, explanation types, use cases, questions, hints, restrictions, verification messages and referral situations. Genialoh provides tutors configured by grade or career, contextualized with the curriculum, values and methodology provided by each institution.

### Module 4. Interpreting reports and analytics

Leaders must learn to distinguish adoption, frequency, usefulness, quality, patterns, academic results and causality. Higher usage frequency does not automatically prove better learning. Reports should be combined with teacher observation, surveys, interviews, activity review, academic evidence and program context.

### Module 5. Improving the strategy

Leader teachers can participate in tutor review, use-case prioritization, rule updates, new training design, program onboarding, incident evaluation, institutional communication and recommendations for leadership.

### Project to complete the level

- Training route for their faculty.
- Activity repository.
- Contextualized tutor.
- Assessment protocol.
- Discipline guide.
- Business case.
- Adoption plan.
- Mentorship system.

## Training modalities

| Modality | Main usefulness |
| --- | --- |
| Talks | Introduce the topic, communicate the strategy, present principles, align messages |
| Workshops | Try tools, design materials, analyze errors, review cases, receive feedback |
| Discipline labs | Adapt uses to programs, goals, methods, assessments and specific risks |
| Microlearning | Short videos, guides, examples, FAQs, weekly cases, demos |
| Consultations | Teachers redesigning assessment, preparing a project or adapting a tutor |
| Communities of practice | Share experiences, errors, materials, activities, strategies and questions |

Talks alone are not enough to develop practical competencies.

## Recommended duration of the full route

| Level | Training hours | Application hours |
| --- | --- | --- |
| Level 1 | 4–6 | 1–2 |
| Level 2 | 6–8 | 3–5 |
| Level 3 | 8–12 | 6–10 |
| Level 4 | 12–20 | 10–20 |

Not every teacher needs to complete the four levels immediately. The institution can require level 1 for all faculty, level 2 for those who will use the platform, level 3 for leads of priority courses and level 4 for coordinators and mentors.

## Eight-week training plan example

| Week | Focus | Content |
| --- | --- | --- |
| 1 | Fundamentals | What AI is, capabilities, limits, verification, first practice |
| 2 | Privacy and responsibility | Data, documents, authorized tools, anonymization, cases |
| 3 | Teaching planning | Objectives, classes, explanations, examples, materials |
| 4 | Activities | Use cases, practice, simulations, discipline design |
| 5 | Academic integrity | Usage levels, disclosures, possible infractions, procedures |
| 6 | Assessment | Process evidence, rubrics, oral defense, verification |
| 7 | Application | Each teacher implements an activity, observes results, records problems |
| 8 | Review | Presentation of experiences, adjustments, recommendations, next level |

## 30-day intensive plan

| Days | Phase | Activities |
| --- | --- | --- |
| 1–5 | Diagnostic and alignment | Survey, interviews, teacher selection, tool definition, communication of goals, rule review |
| 6–10 | Basic level | Fundamentals, limits, verification, privacy, first practices |
| 11–15 | Teaching application | Planning, materials, activities, feedback, discipline cases |
| 16–20 | Design and assessment | Usage levels, redesign, rubrics, disclosures, process evidence |
| 21–25 | Application | Use in an activity, observation, feedback, incident logging |
| 26–30 | Evaluation and continuity | Learning presentation, measurement, mentor selection, adjustments, next-cycle plan |

## How to assess participating teachers

Assessment should not be based on an isolated theoretical exam. It can use evidence such as a contextualized prompt, verification of a response, a reviewed material, a designed activity, a rubric, a use disclosure, reflection, classroom application, results presentation and community participation.

### Teacher competency rubric

| Criterion | Initial | Developing | Competent | Advanced |
| --- | --- | --- | --- | --- |
| Understanding | Recognizes basic concepts | Identifies capabilities and risks | Explains limits and selects proper uses | Guides other teachers |
| Verification | Accepts results with little review | Reviews main aspects | Verifies systematically | Designs discipline protocols |
| Design | Uses general materials | Adapts drafts | Designs aligned activities | Leads curricular redesign |
| Privacy | Recognizes some risks | Anonymizes basic cases | Applies institutional criteria | Coaches complex cases |
| Assessment | Maintains traditional activities | Introduces partial evidence | Redesigns assessment and rubrics | Creates models for programs |
| Leadership | Participates | Shares experiences | Coaches colleagues | Leads communities and projects |

### Certificates and recognition

The institution can recognize participation via attendance certificates, internal badges, level recognition, faculty development hours, community participation, case presentations and mentor roles. Recognition should depend on evidence, not merely on joining a session.

## How to measure the plan's effectiveness

| Category | Indicators |
| --- | --- |
| Participation | Teachers invited, enrolled, attendance, completion, deliverables, participation by faculty |
| Competency | Prompt quality, verification, redesigned activities, rubrics, privacy application, ability to explain limits |
| Application | Implemented use cases, participating courses, materials created, modified activities, tutors used, continuing teachers |
| Usefulness | Teacher perception, perceived preparation time, material quality, difficulties, required support, intent to reuse |
| Safety | Privacy incidents, data doubts, use of unauthorized tools, incorrectly uploaded materials, corrected cases |
| Institutional | Mentors trained, resources shared, participating faculties, onboarded programs, adjustments made, recommendations |

The number of conversations or prompts generated does not by itself prove teaching improvement.

## How to increase adoption

- Start with real problems.
- Work by disciplines.
- Provide an institutional tool.
- Recognize teacher time.
- Offer coaching.
- Show reviewed materials.
- Allow per-course decisions.
- Communicate limits without presenting the tool as infallible.

## Common training mistakes

- Teaching only prompts.
- Offering an isolated talk.
- Training without a common tool.
- Using examples unrelated to disciplines.
- Forcing immediate implementation.
- Not recognizing different levels.
- Ignoring resistance.
- Not providing follow-up support.
- Measuring only attendance.
- Promising automatic time savings.
- Not updating the training.

## Frequently asked questions

### How long should AI training for teachers last?

An introduction can last four to six hours, but a complete route requires practice, application, follow-up and leveled training.

### Should all teachers complete the same modules?

It is advisable to set a common baseline and then adapt training by experience, role and discipline.

### Is it enough to teach how to write prompts?

No. Verification, privacy, academic integrity, activity design, assessment and teacher responsibility must also be covered.

### Do teachers need to know how to code?

Not for most pedagogical use cases. They need to understand goals, context, limits and verification.

### Should all teachers be required to use AI?

Not necessarily. The university can require basic training and let each program determine appropriate uses.

### How is the training assessed?

Through materials, activities, rubrics, classroom application, reflection and verification evidence.

### Can AI be used for grading?

It can support initial organization of criteria or comments, but academic decisions must retain human supervision and responsibility.

### What data must teachers not share?

Names, IDs, records, individual grades, medical or financial information, confidential assessments and unauthorized documents.

### Is it advisable to work with an institutional platform?

It can enable a common experience, curricular contextualization, program tutors, training, support, reports and follow-up.

### Does Genialoh include faculty training?

Yes. Faculty training is part of the institutional process of diagnostic, personalization, training, deployment and follow-up.

### How can a university explore the proposal?

It can start with a 30-minute meeting, select two careers or programs and provide authorized documents to prepare a demo.

## From isolated workshops to institutional capacity

A university does not develop faculty AI capabilities through a single talk. It needs a route that enables understanding, practicing, verifying, designing, applying, assessing, sharing, improving and scaling.

Genialoh enables organizing this training around an institutional platform with the university's name, logo and colors; values and methodology; context from programs and curricula; tutors by career, grade or program; AI support for teachers; personalized learning; reports by student and group; institutional analytics; faculty training; implementation; support and adoption follow-up.

> **Institutional demo**
> 
> Genialoh can prepare a functional demo with the institution's identity and academic context in less than 14 days from receiving the full documentation. After the presentation, up to five authorized people can explore the experience for 30 calendar days.

Book a first 30-minute meeting. During the conversation you will learn about the proposal and select two careers, grades or programs if your institution wishes to move forward.