GenAI in Education

🔬 PTAS-Funded Project: GenAI Integration in Computer Science Education

Principal Investigator: Pavlos Andreadis

Team: Aurora Constantin, Cristina Adriana Alexandru, Fiona Mcneill, Heather Yorston, Judy Robertson, Stuart King, Vidminas Vizgirda; and Anya Habana, Aagoon Chakraborty, Kashvi Chabbra

Institution: School of Informatics, Computer Science Education Group

Funding: Principal's Teaching Award Scheme

Project Motivation

Following the lessons learned from Covid-19 (students needed access to learning materials) and the emergence of ChatGPT (students need learning, not test prep), this project addresses the fundamental shift in how students think, search, and problem-solve with GenAI.

Research Goals

• Understand faculty and student perceptions of GenAI integration
• Document current approaches to GenAI in Computer Science teaching
• Identify challenges and barriers to effective integration
• Develop evidence-based recommendations for educational practice

Research Methodology

Survey Participants:

• 230 Students: 55% male, 36% female, 3% non-binary/third, 6% prefer not to say
• 52 Academic Staff: 50% male, 27% female, 8% non-binary/third, 15% prefer not to say
• Focus Groups: 12 Students, 5 Academic Staff

Key Findings

• Outlawing GenAI is pointless – students will adapt around bans
• Traditional "final-answer" assessment is obsolete – focus on process
• Make students work with/against GenAI – don't ignore or fear it
• Teach prompt engineering and critical AI use – not just memorization
• GenAI can support fundamentals – use it to teach and assess real skills
• AI detectors don't work – don't rely on them
• Exams instead of coursework is only a temporary crutch – don't miss the bigger opportunity

Recommendations

• Lecturers: Start small – experiment with redesigning one assessment
• Administration: Invest in real assessment redesign – this isn't a minor patch
• Get clear on what students should learn and do with GenAI
• Some domains will need GenAI-specific curricula
• Leverage GenAI to enhance teaching and assessment
• The world has changed – our assessment and teaching must change with it

🎓 UKICER 2025 - Conference Organization & Workshop Presentation

Conference: UK and Ireland Computing Education Research (UKICER) 2025

Dates: September 4-5, 2025

Location: Edinburgh, UK

Organizational Role

I am helping organize UKICER 2025 as part of the Local Organising Committee. My main responsibilities include:

• Website Maintenance: Maintaining and updating the conference website
• On-site Administration: Managing logistics and coordination during the conference
• GenAI Workshop: Building, organizing, and delivering an organizer workshop on GenAI in Education

Workshop Presentation

I am excited to be presenting a further GenAI workshop alongside Jake Van Clief, focusing on practical applications of generative AI in educational contexts. This workshop will explore innovative approaches to integrating AI tools in computer science education.

Conference Overview

UKICER is a leading forum for researchers and practitioners to meet and share advances in computer science education. The conference brings together researchers, academics, industry practitioners and teachers from across the UK and Ireland as well as from the rest of Europe and the wider world.

👥 Computer Science Education Group (CSEG) Leadership

Group: Computer Science Education Group (CSEG)

Role: Key Organizer and Active Member

Institution: School of Informatics, University of Edinburgh

Teaching Hours Leadership

I have been a key organizer for dozens of Teaching Hours events, many of which have focused specifically on GenAI in Higher Education. These sessions bring together educators, researchers, and practitioners to share insights, challenges, and best practices in computer science education.

Event Resources

• LinkedIn Group: CSEG LinkedIn Community - Join our professional network
• Events Page: CSEG News and Events - Official events calendar

Note: The old event archive as well as all recent events are not currently accessible, but we aim to recreate at least the comprehensive list of past events.

Group Mission

The Computer Science Education (CSE) group is a platform that brings together peers from the University of Edinburgh, as well as associate members from the UK and beyond, who are involved in teaching and learning and are passionate about Computer Science Education. Our members come from diverse backgrounds and professions, enriching our collective knowledge and enabling us to engage in mutual learning, improve our teaching, foster innovation, and continuously enhance educational practices.

🤖 Learning and Teaching Conference 2025 - GenAI Integration & AI Innovation Panel

Conference: University of Edinburgh Learning and Teaching Conference 2025

Presentation: PTAS Project: GenAI Integration in Computer Science Education

Panel: AI for Teaching Innovation Panel Member

Institution: School of Informatics, University of Edinburgh

PTAS Project Presentation

I presented our comprehensive PTAS-funded research on GenAI integration in Computer Science Education, sharing the key findings from our survey of 230 students and 52 academic staff, along with evidence-based recommendations for educational practice.

AI Innovation Panel Contributions

As a panel member on the "AI for Teaching Innovation" discussion, I contributed insights on:

• Co-Mod Project: Discussed our work towards a Feedback Moderator and Coordinator system
• Environmental Impact Debate: Challenged prevalent misconceptions about GenAI's environmental cost
• Training vs. Inference: Drew clear distinctions between training and inference time costs
• Mature Discussion Framework: Advocated for more nuanced understanding of AI's environmental impact

Environmental Impact Discussion

I addressed the critical issue of how naive handling of GenAI's environmental impact by colleagues has hindered student development. The PTAS survey results clearly show that overly cautious approaches based on misinformation have prevented students from gaining essential AI literacy and skills.

Key Arguments Presented

• Inference vs. Training: Emphasized that inference time (what students use) has minimal environmental impact compared to training
• Educational Opportunity Cost: Highlighted how avoiding GenAI due to environmental concerns deprives students of essential skills
• Evidence-Based Approach: Advocated for data-driven decisions rather than fear-based restrictions
• Student Development: Stressed the importance of preparing students for an AI-integrated workforce

Impact

This presentation and panel participation helped shift the conversation from fear-based restrictions to evidence-based integration strategies, contributing to a more mature and productive discussion about GenAI in education at the University of Edinburgh.

📊 Learning and Teaching Conference 2023 - Learning Analytics at Scale

Conference: University of Edinburgh Learning and Teaching Conference 2023

Publication: Designing Useful Learning Analytics: Developing an Adaptable LA Dashboard

Authors: Paul Boocock, Stefania Sandu, Xinyi Liu, Qiwen Liang, Heather Yorston, Aurora Constantin, Cristina Alexandru, Pavlos Andreadis

Institution: School of Informatics, University of Edinburgh

Research Focus

This work addressed the critical need for systematic Learning Analytics (LA) infrastructure rather than ad-hoc solutions. The research focused on designing for co-design of course data and interventions, rather than designing the LA experience in an ad-hoc basis.

Key Innovation

Our approach emphasized the development of an adaptable LA Dashboard that works School-wide and adapts to the requirements of individual users, computing infrastructure, and data availability. This represented a shift from custom, one-off solutions to a scalable, systematic framework.

Methodology

• Cross-Project Collaboration: Worked across multiple taught student projects
• Infrastructure Development: Created adaptable dashboard architecture
• User-Centered Design: Focused on individual user requirements
• Data Integration: Addressed varying data availability scenarios

Connection to GenAI Work

This project served as a proving ground for the systematic approach I now advocate for in GenAI integration in Education. The lessons learned about designing scalable, user-adaptive systems directly informed my current work on GenAI integration frameworks and assessment redesign.

Impact

The research identified key barriers to LA adoption and provided a foundation for systematic, scalable learning analytics implementation that can be adapted to various educational contexts and technologies.

🎪 Informatics Teaching Festival 2022 - Design of Teaching and Learning

Event: Informatics Teaching Festival 2022

Dates: May 9-11, 2022

Theme: Design of Teaching and Learning

Institution: School of Informatics, University of Edinburgh

My Role

I served as a co-organizer and session chair for this teaching festival, a milestone for the School of Informatics, which focused on the design of teaching and learning in computer science education. The festival brought together educators, researchers, and practitioners to share innovative approaches and best practices.

My Contributions

• Co-Organizer: Helped plan and coordinate the three-day festival structure
• Session Chair: Facilitated discussions and managed session flow
• Presenter: Delivered a presentation in one of the sessions
• Panel Leader: Led a panel discussion on the employability of our graduates

Festival Structure

• Day 1: Overview of Course Design
• Day 2: Design to Develop Student Skills, including for Industry
• Day 3: Design of Assessment

Key Focus Areas

The festival addressed critical aspects of modern computer science education, including:

• Process and experience of designing new courses
• Support for course design through ELDeRs (Experienced Learning Designers)
• Sharing positive experiences on improved courses
• Assessment design and marking approaches
• Rubric-based marking schemes
• Challenging assignments without encouraging excessive time investment

Impact

This festival represented a significant milestone in advancing teaching excellence within the School of Informatics, providing a platform for educators to learn from each other and develop innovative approaches to computer science education.