Budapest University of Technology and Economics (BME) and Óbuda University are introducing artificial intelligence (AI) into teaching, learning and administrative processes to make education more accessible and personalized. Both institutions emphasize that while AI can augment education, human responsibility and critical thinking must remain central.
BME: validation, engineering responsibility and educational tools
At the BME Faculty of Civil Engineering the role of the human decision-maker is particularly prominent because choices about sizing, load-bearing capacity and durability of steel structures can affect human lives. Dr. Joó Attila László, associate professor at the Department of Bridges and Structures, described such design as a “hazardous operation” where final responsibility cannot be shifted to AI.
The faculty received an invitation from the European Committee for Standardization to join a working group that will propose provisions for AI-based design of steel structural members in the next generation of Eurocode 3 (EN 1993). The aim is to make AI-supported design transparent, explainable, reproducible and able to be validated under defined conditions.
Across Europe hundreds or even thousands of AI agents are being developed for use in education and construction design. The BME, which conducts leading research on AI applications in construction, could play a significant future role in validating these systems because enforcing safety and transparency is an urgent need.
Within an NKFIH Market Focus Area grant, the university developed a specialised AI agent in cooperation with Neuron Solutions Kft., built on the Microsoft Azure platform. This agent is fed by several thousand pages of the faculty’s lecture and textbook materials and is intended to serve as a basis for validating other AI agents.
At BME, AI is used extensively in research and teaching: it accelerates model building, performs large-scale data computations, and makes new theories easier to illustrate. The university has also developed AI solutions to make learning more engaging and personalised, and to automate some assessments that were previously only reviewable by humans—for example, AI-supported evaluation of engineering drawings that provides interactive exercises and immediate feedback to deepen students’ skills.
Dr. Joó stated: “Ours is the last generation that experienced making mistakes as humans, and we learned a lot from that. Even in the AI era this knowledge will be needed… It is important to realize that final responsibility always rests with humans.” He also noted the risk that AI use can be isolating, and argued for team-based education and for academic teachers to take on more facilitative and mentoring roles to ensure safe collaborative learning and quality of education.
Óbuda: AI Campus and an operational model shift
Óbuda University treats AI as a driver of an institutional transformation rather than as a set of purchased tools. Increasing numbers of instructors automate workflows and use Copilot Studio to develop custom AI agents and chatbots. HelpMate, available within Microsoft Teams, is one such application: it gives immediate answers to administrative and student queries and supports study administration.
The university’s AI transformation is coordinated by the AI Campus, which aligns initiatives, builds an operational model from proven solutions, develops competencies and measures AI’s impact on teaching and institutional operations. Prof. Dr. Eigner György, dean of the Neumann János Faculty of Informatics at Óbuda University, said: “A university does not become AI-native by buying many AI tools, but by simultaneously transforming its own operations and developing the capabilities of the people who use AI.”
The university provides hybrid, platform-independent infrastructure, training, support and an appropriate organisational culture, while continuously assessing where AI can add value and what risks it may bring. Eigner emphasised that instructors are expected not merely to use AI tools but to actively shape the pedagogy of the AI era, including new forms of teaching and assessment.
Limits of educational AI and the need for awareness
International data indicate high use of AI among students but limited formal training. Microsoft research finds that 92% of school and higher education students use AI for school tasks, while 77% have not received formal instruction on how to use AI.
In May this year Microsoft released the Study and Learn Copilot agent, designed specifically to support education: rather than offering quick, ready-made answers, it encourages what the company calls productive struggle to support active learning. Microsoft’s view is that responsible AI use goes beyond well-crafted prompts: it involves understanding how the technology works and its limitations, and—above all—the capacity for independent judgment.
Both universities agree that in a few years AI will not be an isolated digital tool but an integral part of learning, research and work. Preparing people and organisations for responsible, value-generating AI use therefore requires institutional culture, infrastructure and training that preserve critical thinking and human accountability.
Conclusion
The BME and Óbuda University examples show that integrating AI into university life is a technical, pedagogical and ethical project at once. Alongside developing and deploying tools, universities must provide training, design validation procedures, and build institutional practices that ensure AI supports—but does not replace—human responsibility and independent decision-making.
