Engineering institutions in India are rethinking how they teach, assess, and prepare students as artificial intelligence reshapes both the classroom and the workplace. Anubha Gupta, Head of ECE and Professor at IIIT Delhi, argues that the shift requires stronger fundamentals, not weaker ones, and that engineers must learn to question what AI produces rather than simply accept it.
"AI should not replace engineering education, but it should strengthen it," Gupta said. "While generative AI can write code and summarise knowledge, it cannot judge whether a solution is correct, safe or appropriate for a given context."
Teaching with AI, not despite it
Gupta said courses should move from content delivery to problem-driven learning, with each subject ending in an application-oriented interdisciplinary project. AI should act as a learning partner, not a substitute for understanding. The goal, she said, is to produce engineers who can reason, evaluate AI-generated outputs, and make informed decisions.
The same logic applies to assessment. Gupta said institutions should treat generative AI as an opportunity rather than a threat. Students can use it to brainstorm, explore alternatives, and refine designs. Exams, in turn, should test reasoning, creativity, and teamwork through design reviews, oral examinations, laboratory work, and open-ended projects rather than information recall.
"Academic integrity should emphasise transparency and responsible AI use rather than prohibition," she said. "The educational value of AI lies not in answering questions but in helping students ask better ones and think more critically."
Interdisciplinary skills for real-world problems
Healthcare engineering shows why interdisciplinary training matters, Gupta said. Building effective AI tools for medicine requires understanding clinical workflows, medical data, ethics, and regulation before choosing an algorithm. The same applies to agriculture, climate science, and manufacturing.
Engineering schools should respond with domain-specific minors, collaborative capstone projects, and partnerships with hospitals, industries, and public agencies. "The future engineer must be fluent not only in engineering but also in the language of the application domain," she said.
This extends to research. Gupta said engineering research should go beyond publications to include prototypes, patents, startups, and deployable solutions. Students should start with real-world problems rather than hunting for uses of popular technologies. Institutions can support this through group projects and by recognising societal impact alongside academic output. For educators adapting to these expectations, structured training in AI Learning Path for Teachers offers practical methods for integrating AI into instruction and assessment.
Rethinking the degree structure
Gupta said India's National Education Policy already provides a strong base through multidisciplinary education, flexible curricula, and experiential learning. The next step is implementation. She proposed moving beyond the current Choice-Based Credit System toward learner-designed academic pathways, where students combine courses, research, and internships across institutions to build competencies aligned with emerging fields.
"Future engineering education should be competency-driven rather than constrained by traditional departmental boundaries," she said.
Graduates, she added, will need a mix of technical and human skills: AI literacy, data science, cybersecurity, and systems engineering, alongside critical thinking, communication, and adaptability. The capacity for lifelong learning, she said, will become one of the most valuable professional competencies. For professionals and faculty seeking structured ways to build these capabilities, AI for Education resources cover curriculum integration, responsible AI use, and assessment redesign.
Why this matters for educators
For faculty and administrators, the practical takeaway is direct: update how you evaluate students before they update how they work. If assignments can be completed by a chatbot, the assignment needs revision. Design assessments that require reasoning, experimentation, and presentation. Teach students when to trust AI output and when to challenge it. That shift, Gupta said, is what separates an engineer who uses tools from one who understands them.
Your membership also unlocks:
