Learning and assessing smarter with AI
How do we responsibly integrate AI into an existing curriculum? How do we safeguard educational standards when using AI? And in what ways can AI enhance learning and assessment while preserving essential human expertise? In this webinar Carola Strobl (UAntwerpen) discusses the policy-related and pedagogical approach to embedding AI literacy in a language-focused programme. Arne Vanhoyweghen (VUB) showcases a practical implementation of AI in teaching that assists teachers in grading handwritten math tests more efficiently without compromising quality.
Learn from colleagues how they use AI tools to support and adapt their daily work, education, and research.
Practical information:
Want to register?
- Price: €0
Part of Peers & Prompts, online AI inspiration sessions for higher education, an initiative of the VLIR working group "Generative AI in university education," in collaboration with VAIA.
Georganiseerd door:
Leertraject
AI-geletterdheid in de Bachelor Toegepaste Taalkunde: een beleidsmatige en didactische aanpak
Carola Strobl, UAntwerpen
Carola Strobl belicht het spanningsveld bij de integratie van AI-geletterdheid in een taalgericht studieprogramma, meer bepaald de Bachelor Toegepaste Taalkunde aan de UAntwerpen. AI wordt in de publieke opinie beschouwd als dé grote concurrent voor taalprofessionals die in dit programma worden opgeleid; de studenten krijgen dus vaak de vraag te horen waarom het überhaupt nog zinvol is om talen te studeren. De opleidingscommissie heeft daarom snel geschakeld en al in 2023 een werkgroep Digitalisering en AI in het leven geroepen. Carola licht de werking van deze groep toe, die tot doel heeft AI-competenties in de bestaande leerlijnen van het studieprogramma te integreren en alle stakeholders – studenten, docenten en het potentiële werkveld – hierbij te betrekken. De werking van de groep steunt op drie pijlers: inventariseren (van ervaringen en van onderzoeksresultaten), adviseren (door middel van richtlijnen en eindcompetenties) en connecteren (door de organisatie van uitwisselingssessies). Bij de stapsgewijze integratie van AI in het curriculum wordt een didactisch verantwoorde balans nagestreefd tussen het opbouwen van eigen vaardigheden en kennis zonder hulpmiddelen enerzijds en de integratie van nieuwe vaardigheden in samenwerking met AI anderzijds: “omarmen, niet bannen”.
In het Nederlands!
AI-Assisted Grading of Handwritten Math Tests
Arne Vanhoyweghen, VUB
Giving students quick, personal feedback on handwritten exams is one of the best ways to support learning, but it takes a great deal of time. At the same time, the rise of generative AI has made traditional take-home assignments less reliable, increasing the need for supervised, in-class assessments that still require careful grading.
In this talk, Arne Vanhoyweghen presents a practical system that uses AI to assist teachers in grading short, handwritten tests. The approach combines clear grading guidelines with AI support, automated scanning, and multiple independent evaluations of each answer. Importantly, every grade is still reviewed by a human instructor to ensure fairness and accuracy.
We tested this system in two undergraduate mathematics courses across six in-class assessments. The results show that AI assistance reduced grading time by about 23%, while maintaining grading quality comparable to fully manual correction. Although AI occasionally makes mistakes, the built-in human oversight effectively prevents these from affecting final grades.
Overall, the results show that AI can meaningfully accelerate grading without lowering standards, provided it is embedded in a carefully designed human-in-the-loop workflow.
In English!
Moderator
Saskia Goeteyn, VAIA
Facilitates the session and facilitates discussion between speakers and the audience. Saskia will ensure the session runs smoothly.
Teachers / speakers
Carola Strobl
Carola Strobl is hoofddocent aan het Departement Toegepaste Taalkunde, Vertaal- en Tolkwetenschap van de Universiteit Antwerpen. Ze doceert Toegepaste Taalkunde en Vertalen Duits-Nederlands en is voorzitter van de werkgroep Digitalisering en AI binnen de onderwijscommissie. In haar onderzoek spitst zij zich toe op vreemdetaal- en vertaaldidactiek, met een bijzondere focus op het gebruik van digitale tools in het leerproces. Een tweede interessegebied is corpusgebaseerd onderzoek naar kenmerken van leerderstaal en vertaalde taal.Â
Arne Vanhoyweghen
Arne Vanhoyweghen is an AI researcher and data scientist working to make large language models more reliable, interpretable, and deployable in high-stakes settings. He is a postdoctoral researcher at the Data Analytics Lab of Vrije Universiteit Brussel and a guest professor in Business Technology and Operations at the Faculty of Social Sciences & Solvay Business School.
His research investigates how LLMs reason, where they fail, and how internal signals in their outputs can be harnessed to detect uncertainty and improve calibration. He applies these insights to design structured AI workflows for education, encompassing automated assessment, feedback generation, and scalable learning assistance. These systems integrate advanced prompting strategies, multi-pass evaluation, and human-in-the-loop verification to ensure reliability, fairness, and pedagogical alignment.
His expertise spans LLM evaluation, uncertainty detection, chain-of-thought analysis, AI-assisted assessment, and human–AI collaboration systems. In parallel, he conducts research in Ergodicity Economics, where he studies decision-making under non-ergodic dynamics and long-run growth processes, linking formal economic theory with computational modeling.
Peers & Prompts: AI Inspiration for Higher Education
The most surprising discoveries often happen close to home. This also applies to generative AI! Are you curious about how your colleagues are using AI tools to innovate education and research? Want to learn how to integrate (gen)AI into your daily work? Then join our monthly online sessions and discover inspiring practices from Flemish university education.
Learn from colleagues how they use AI tools to support and adapt their daily work, education, and research.
Related courses
Fundamental concepts of multi-omics data integration
Opleiding - Leuven - Genomics Core Leuven
