June 2026
Training the Teachers for the Age of AI
Why faculty development must move beyond tool workshops
By Adarsh Lathika, Founder, Anatomy of Work
Students are already using AI.
Some use it to simplify difficult readings. Some use it to prepare for viva questions. Some use it to polish assignments. Some ask it to explain concepts in simpler language. A few use it as a thinking partner. Most use it as a shortcut.
The question before higher education is no longer whether students will use AI. They already are. The more important question is: how should faculty now teach because AI exists?
For the past several months, as part of developing the Anatomy of Work Teaching Framework, we have been speaking with educators, institutional leaders, students, founders, and workplace learning teams around the world, including India. Across these conversations, one pattern has become clear: AI is not entering education through policy alone. It is entering through everyday behaviour.
Students use AI quietly to decode concepts, prepare assignments, rehearse answers, and improve the polish of their work. Faculty notice that submissions are becoming more fluent, but not always more thoughtful. Institutions are setting up AI initiatives, but many are still asking what should change in the actual classroom. Employers increasingly expect graduates to work with AI, but assessments still often reward memory, fluency, and final output more than reasoning, judgment, and application.
This is why training teachers for the age of AI cannot be limited to a one-time workshop on tools. Faculty need support to redesign how learning happens, how assessment works, how students use AI, and how human capabilities are built in an AI-shaped world.
The hidden AI classroom
In several Indian higher education conversations, faculty members have described a familiar discomfort: assignments have become cleaner, more structured, and more fluent, but not always more thoughtful. Some students can submit a polished answer but struggle to explain the reasoning behind it when asked orally. Others use AI to generate presentation outlines or project summaries, but cannot defend the assumptions behind the work.
This is not simply a cheating problem. It is a visibility problem.
Much of student AI use is happening quietly. Students may not disclose that they used AI to summarize a reading, structure an essay, create a slide deck, debug code, write an email, or prepare for an interview. In effect, a hidden curriculum is forming. Students who know how to ask better questions, verify outputs, rewrite prompts, and combine AI with their own judgment gain an invisible advantage.
In India, this matters even more because classrooms are deeply unequal. A student from an English-medium, urban, digitally confident background may use AI very differently from a first-generation learner who is still building academic confidence. A student with paid AI tools may experience a different learning environment from a student depending on free tools and patchy access.
If faculty ignore AI, this inequality may grow. If faculty bring AI into the open, it can become a learning opportunity.
Faculty training must therefore begin with a simple recognition: AI use is already happening. The task is not merely to permit or prohibit it. The task is to make AI use visible, discussable, ethical, and learnable.
Why faculty anxiety is valid
Faculty concerns should not be dismissed. There are real risks.
AI can weaken effort if students outsource too early. It can make students appear more capable than they are, because the final answer may be fluent while the underlying understanding is weak. The larger problem is not only that AI can produce incorrect answers, but that students may not have the ability, patience, or habit to validate what they copy. When questioned, they may struggle to explain the logic, assumptions, or evidence behind their own submission.
AI can also reduce the productive struggle that is necessary for learning. It can make writing look better than thinking. It can create unfairness between students who use AI responsibly and those who use it invisibly. It can blur the line between assistance and authorship.
In one Indian conversation, a faculty member described the central dilemma sharply: if the assignment is generic enough for AI to answer, was it ever a good assessment of student capability?
That question deserves attention.
AI is exposing something uncomfortable. Many traditional assessments were already vulnerable because they rewarded reproduction, formatting, and fluency more than judgment, reasoning, and application. AI has not created this weakness; it has made it visible.
The response cannot be only detection. AI detection tools are imperfect, and a classroom built around suspicion can damage trust. The deeper response is teacher preparation. Faculty need to be trained not only to identify misuse, but to redesign learning so that students must demonstrate reasoning, judgment, and understanding.
What faculty need to be trained for
Faculty training must go beyond “how to use ChatGPT.” It must prepare teachers to make better pedagogical decisions in an AI-shaped learning environment. Six areas are especially important.
1. AI literacy beyond tools
AI literacy cannot be reduced to prompt engineering. It should also be tool-agnostic.
The specific tools will keep changing. Students may use ChatGPT today, Gemini tomorrow, Copilot in the workplace, and domain-specific AI systems in their future professions. If AI literacy is taught only as tool usage, it will become outdated quickly.
Faculty need to understand what AI can do, where it fails, why it may produce fluent but unreliable answers, how bias can enter outputs, how hallucinations occur, and how to model responsible use for students. The goal is not to master one platform. The goal is to understand how to teach responsibly when AI tools keep changing.
2. Assignment and assessment redesign
AI breaks weak assessments first.
Generic essays, predictable questions, standard summaries, and take-home assignments that reward polished output are now easier to complete with minimal understanding. This does not mean assignments should disappear. It means assignments must be redesigned.
Faculty need training on how to create tasks that require context, judgment, and explanation. This may include local case studies, live problems, field observations, oral defenses, iterative drafts, reflection notes, critique-based tasks, and assignments where students compare their own thinking with AI-generated responses.
A polished answer no longer proves learning. Faculty need to assess reasoning trails, drafts, source validation, student reflections, and oral explanations. They need rubrics that reward clarity of thought, quality of evidence, ability to critique, and the student’s capacity to explain and defend their work.
The shift is from asking students to only submit an answer to asking them to show how they arrived at the answer.
3. Teaching students to use AI for learning
Students often use AI as a shortcut. Faculty need to be trained to convert AI use into a learning loop.
This means helping students ask better questions, compare answers, check sources, identify assumptions, revise drafts, and reflect on what they learned. It also means teaching students when not to use AI, when to struggle with a problem first, and when AI can be used as a scaffold rather than a substitute.
In one AoW-linked discussion on writing and thinking, an important idea emerged: writing should not be treated only as a final product. Writing is a way of thinking. If AI produces the first draft, the faculty task is to examine what the student did next. Did the student question it? Improve it? Localize it? Add evidence? Challenge its assumptions? Connect it to lived experience or field observation?
Faculty cannot merely say “use AI responsibly.” They need to demonstrate what responsible use looks like in their subject.
4. Discipline-specific AI pedagogy
AI will not be used in the same way across disciplines.
A commerce faculty member, an engineering faculty member, a humanities faculty member, a law faculty member, a medical faculty member, and a management faculty member will each face different questions. What counts as responsible AI use in coding may not be the same as responsible AI use in clinical reasoning, legal analysis, design critique, financial modelling, or literary interpretation.
This is why generic AI workshops rarely change classroom practice. Faculty need examples from their own disciplines.
Training should help faculty ask: What does responsible AI use look like in my subject? What should students be allowed to automate? What must they still learn deeply? What kinds of tasks should they be able to explain without AI? What kinds of AI use should they disclose?
5. Feedback and workload management
It is easy to ask faculty to redesign learning. It is harder to acknowledge the workload they already carry.
Many faculty members are managing teaching hours, administrative responsibilities, evaluation, accreditation documentation, mentoring, placement support, institutional reporting, and often research expectations. If AI is introduced as yet another responsibility, it will create resistance. If AI helps reclaim faculty time, it can create room for genuine teaching innovation.
One of the first uses of AI in higher education should therefore be to reduce repetitive academic labour.
Faculty can use AI to generate practice questions, create rubrics, prepare caselets, draft feedback templates, summarize common student errors, create viva question banks, design simulations, and produce multiple explanations for the same concept at different levels of difficulty. AI can also help faculty convert one teaching idea into several learning formats: a classroom discussion, a quiz, a reflection prompt, a project brief, and an assessment rubric.
The goal is not to automate the teacher. The goal is to free the teacher for higher-value work: judgment, mentoring, discussion, contextualization, assessment design, and intellectual challenge.
6. Ethics, disclosure, and learning science
Students need clarity. In many classrooms, they do not know what counts as acceptable AI use. Can they use AI to brainstorm? Summarize a reading? Improve grammar? Draft an answer? Should they disclose it?
Faculty need training on how to set clear norms: what is allowed, what is limited, what is prohibited, and what must be disclosed. Ambiguity creates unfairness. Some students may avoid AI completely out of fear. Others may use it invisibly. Some may use it as a learning aid, while others may submit AI-generated work as their own.
Faculty should also be supported with basic learning science: attention, memory, retrieval, productive struggle, feedback, revision, metacognition, and transfer. These ideas matter because AI can either support or bypass learning depending on how it is used.
If a student asks AI for an answer before attempting a problem, learning may weaken. If the student first struggles, then uses AI to compare approaches, test assumptions, and revise thinking, learning may deepen.
Faculty training must therefore include not only how AI works, but how learning works.
Why continued faculty development matters
Teacher training for AI cannot be a one-day faculty development programme.
AI tools will change. Student behaviour will change. Workplace expectations will change. Assessment vulnerabilities will change. Institutional policies will evolve. New evidence will emerge on what helps or harms learning.
Faculty development must therefore become a continuing professional practice.
Institutions can support this through short monthly faculty circles, discipline-wise sharing sessions, peer review of AI-aware assignments, industry interactions, learning science sessions, student feedback loops, repositories of working examples, and annual refreshers on policy and tools.
The purpose is not to make every faculty member an AI expert. The purpose is to help every faculty member become confident enough to make sound pedagogical decisions in an AI-shaped learning environment.
Faculty cannot do this alone
If faculty are expected to redesign assessment, support diverse learners, teach AI literacy, prepare students for uncertain careers, and use AI responsibly themselves, then the transition cannot rest on individual effort alone.
An AI-shaped classroom requires institutional support. This need not always begin with expensive platforms. Institutions can start with shared guidelines, faculty peer circles, common assessment templates, AI-use disclosure norms, sample rubrics, and small pilots within departments.
If institutions want faculty to move from content delivery to learning design, they must give faculty time, legitimacy, and support to make that transition.
Institutions can begin with low-cost but high-impact steps: study global practices in AI-enabled teaching and assessment; invite industry experts to explain how AI is changing work; involve learning science experts to ensure that AI use strengthens, rather than weakens, cognition; create department-level AI teaching circles; build a shared repository of AI-aware assignments and rubrics; encourage faculty to document small classroom experiments; recognize teaching innovation; create student AI mentors or teaching assistants; and develop clear norms on disclosure, citation, and acceptable use.
The transition to AI-enabled learning cannot be faculty-led in responsibility but institution-neglected in support.
A practical teacher-training playbook
A useful starting point could be a 5C playbook for faculty development.
Clarify. Train faculty to define what AI use is allowed, limited, or prohibited in their courses.
Contextualize. Train faculty to design tasks rooted in local examples, field observations, current events, personal reflection, or live organizational problems.
Critique. Train faculty to ask students to evaluate AI-generated answers: What is missing? What is inaccurate? What assumptions are hidden? What would change in the Indian context?
Co-create. Train faculty to use AI to generate examples, simulations, role plays, counterarguments, quizzes, and feedback prompts.
Certify thinking. Train faculty to assess reasoning, judgment, application, and the ability to explain one’s process.
The India opportunity
India has an opportunity to develop a distinctive model of AI-enabled education.
We should not simply import Western AI policies into Indian classrooms. Our context includes large classrooms, linguistic diversity, uneven access, employability pressures, public-private variation, and first-generation learners. The Indian AI-in-education conversation must therefore be both ambitious and grounded.
AI can widen inequality if only confident students use it well. But it can also reduce inequality if faculty intentionally use it to provide scaffolding, practice, feedback, translation, and confidence-building.
AI can weaken learning if it becomes a shortcut. But it can deepen learning if students are taught to question, revise, verify, and apply.
The future of faculty is not less human. It is more deliberately human.
But faculty need time, training, tools, institutional support, and students who are gradually taught how to learn.
The real question is not whether AI can teach. The real question is whether we can train and support teachers to redesign learning for a world where intelligence is abundant, but judgment remains scarce.
That may be one of the most important educational tasks before India today.
