AI in Education: Practical insights from ETH teaching staff
By Karin Brown and Daniel Flück, with assistance by generative AI*. As generative artificial intelligence (GenAI) continues to reshape education, educators are discovering both the promises and pitfalls of integrating AI tools into teaching. Our recent analysis following 4 workshops with a total of 48 participants reveals crucial insights about effectively implementing AI in educational settings while maintaining pedagogical integrity.
Prior to the workshop, participants engaged in a structured online preparation phase. This step asked them to experiment with GenAI tools in their subject areas and identify potential implications for their courses. This hands-on experience ensured that our discussions in the workshop were grounded in practical application rather than theoretical possibilities.
The Power of Precise Prompting
One clear lesson emerges: the quality of GenAI output directly correlates with the quality of our prompts. Generic prompts yield generic answers, while specific, well-structured prompts produce valuable results. Educators are finding success by treating prompting as a skill to be mastered – breaking down complex questions, providing clear context, and iterating when necessary.
Finding the Sweet Spot
GenAI isn’t magic, but it’s proving incredibly useful in specific areas. It excels at:
Streamlining administrative tasks and generating initial drafts and course structures. It can provide quick summaries of complex materials, offer step-by-step explanations and support personalised learning approaches.
However, it’s important to recognise its limitations. GenAI struggles with technical diagrams, truly creative work, and tasks requiring deep contextual understanding. It’s a tool to augment teaching, not replace critical thinking.
Transforming Teaching Methods
The integration of GenAI is gradually shifting traditional teaching paradigms. We’re seeing potential for reduced reliance on frontal teaching, more interactive, personalized learning experiences, enhanced focus on critical evaluation skills, new approaches to student assessment and feedback as well as opportunities for more engaging course content.
Critical Considerations
As we embrace these tools, several important considerations emerge: Students need training in verifying GenAI outputs as they may lack the expertise to judge results. Copyright, intellectual property and ethical issues require careful attention. It’s necessary to invest time in learning how to use GenAI tools. Possible societal impacts need ongoing evaluation and maintaining a human connection in teaching remains central.
Looking Forward
The key to successful GenAI integration lies not in wholesale adoption but in thoughtful implementation. Educators should:
1. Start with clear learning objectives
2. Maintain a critical mindset
3. Focus on developing students› GenAI literacy
4. Use AI to enhance, not replace, human interaction
5. Stay updated with evolving capabilities and best practices
Conclusion
GenAI in education isn’t about replacing teachers or traditional methods – it’s about enhancing our ability to educate effectively. One participant said, it’s like using an e-bike; It’s you, just faster! By understanding both its capabilities and limitations, we can harness GenAI’s potential while maintaining the crucial human elements of education. The goal isn’t to create GenAI-dependent classrooms but to use GenAI as one of many tools in our educational toolkit.
If there was one message the workshop leaders wanted to pass on to the participants, it’s the importance of talking about GenAI and it’s use in our academic and scientific lives.
The future of education will likely be a blend of traditional teaching wisdom and AI-enhanced capabilities. The key is finding the right balance for each educational context while keeping student learning and development at the center of our decisions.
New dates for workshops have been published on the UTL Website. Find out how to register here.
*Generative AI was used to transcribe photos of posters participants created with “lessons learned and recommendations”. These transcriptions were then summarised for a first draft of the themes mentioned in this blog.