Artificial Intelligence: Possibilities and Challenges abstracts for the Language centre days 2023
Katrin Korkalainen and Eva Braidwood, University of Oulu
Eyeing the AI revolution: supporting learners’ understanding of the use of ChatGPT
During the academic year 2022-2023, we witnessed the onset of a technical revolution in the field of artificial intelligence (AI), which instigated a new turn in the world of education. In particular, the arrival of ChatGPT, a large language model-based chatbot, shook the very foundations of language education and forced academics and educators to reassess their approaches to teaching and learning, and reconsider, especially, writing-based assignments. While teachers’ reactions were instantaneous and diverse (ranging from outcries lamenting the ‘death of education’ to voices embracing AI as an instrument enabling new methods), little was yet known about students’ perspectives, particularly in Finland, on the uses of AI in higher education.
While waiting for institutional guidelines concerning AI, and having recognized the importance of learners’ opinion, we designed a small-scale action-research study to gauge the student perspective on the matter. Thus, in spring 2023, under the momentum of the AI revolution, we created an experimental task in the “English for Architecture” course (Oulu University). The aim was to test the use of ChatGPT in a writing assignment and to investigate students’ attitudes and understanding of the tool. For this purpose, we designed a questionnaire, which meant to shed light on their views of using AI in higher education and its suitability for learning. The continuation of this project in the “English for Urban Design & Planning” course in autumn 2023 with the same group of students allowed us to further experiment with various AI tools to detect any shifts in attitude or aptitude regarding the usage of AI for learning. In the end, the findings of this small-scale research will be used for developing these courses and design tasks to benefit students both in their learning and their future work.
Johannes Sumuvuori, University of Vaasa
Learning and Teaching with AI: Get It Done Faster and Better
”Workers who work with generative AI will replace those who don’t” – Erik Brynjolfsson in an interview to CNBC, 2023.
In recent years, the landscape of communication and education has undergone a seismic shift, prompting a reevaluation of traditional methodologies and the integration of innovative technologies. Through this presentation, attendees will gain practical insight into harnessing Artificial Intelligence (AI) to make both their own and their students’ work more efficient and effective to meet the game-changing gains reported in the private sector, while also preparing them to the critical challenges.
We will discuss the multifaceted possibilities Large Language Models (LLMs) present in language education through three practical cases, including personalized learning paths and real-time feedback, while also acknowledging the problems these tools pose, including ethical and security considerations, and the potential for inequality in access.
A significant focus will be on how the integration of LLMs in language learning relates to working life, especially in developing communication and research skills vital in today’s accelerating and globalized workplace. We will examine the evolving needs of the workforce and how AI-driven workflow can effectively meet these demands.
Additionally, the presentation will address the crucial aspect of assessing learning in an LLM-augmented environment. We will explore suitable assessment methods, both old and innovative, that not only evaluate language proficiency but also adapt to individual learners’ progress, offering more relevant and comprehensive measuring of skills.
The presentation will illustrate these points through three practical case studies, eschewing the technical intricacies of AI development to focus on the educators’ and learners’ perspective. These cases will demonstrate how AI has revolutionized language learning, communication, and work, necessitating a paradigm shift from educators.
This abstract and the associated summary were written with the help of ChatGPT 4 and LanguageTool. See the ChatGPT thread.
Maria S Lampinen, Jenny H Tarvainen, Christopher Jarvis, University of Jyväskylä
Designing a pilot AI literacies course
Publicity around ChatGPT and the use of large language models since 2022 has increased demand on universities for pedagogical responses that build students’ capacity and resilience, mitigate a widely discussed range of risks while maximising the opportunities offered by these technologies as solutions to problems that are still being explored (or created). In some cases, the risks and problems are understudied or unevenly articulated. Meanwhile, the technology itself is changing.
In the Centre for Multilingual Academic Communication at the University of Jyväskylä, we attempt to hit this moving target by piloting a 3-4 ECTS course to teach ‘AI literacies’ to second-fifth year university students from all disciplines. Our surveys of teacher colleagues revealed a set of ideas in tension around, e.g., learner autonomy versus gatekeeping, and what are the relevant skills to teach in a changing world. How do students’ expectations map onto these tensions, and how have we negotiated our own assumptions and interests as a multilingual and multicultural team, while managing ourselves as a resource? We chose to structure the course around research and writing processes, leaving as much space as possible for students to develop interdisciplinary AI literacies through co-creation and experiment. We discuss the brief, how we approached it, the challenges and how we hope to address them. We invite a discussion about the issues.
Susanna Hakulinen-Bernard, Miro Metsämuuronen and Henriikka Rinta, Tampere University
Kirjallisen kielitaidon arvioinnin nykyisyys ja tulevaisuus
Arviointikäytänteet ja niiden mukauttaminen uusiin tekoälymallien tuomiin haasteisiin puhuttavat korkeakouluissa. Myös meillä Tampereen yliopiston kielikeskuksen ruotsin tiimissä on v. 2022–23 haettu uutta linjaa kirjallisen taidon arviointiin, ja haluamme keskustella aiheesta kielikeskuskontekstissa laajemmin. Haasteita tuovat kielimallien lisäksi paine saada opiskelijat valmistumaan nopeammin ja toisaalta opiskelijoiden jännitys, motivaation puute tai heikohko lähtötaso. Erityisen haasteellisena moni opettaja kokenee myös ns. monenlaisten oppijoiden tilanteen digimaailmassa. Yhä useampi yliopisto-opiskelija tarvitsee yksilöllisiä opiskelujärjestelyjä, mikä voi aiheuttaa ristiriitoja arviointikäytänteiden osalta.
Niemisen (2021) mukaan arviointi korkeakouluissa käsitetään yhä tenttien ja suorittamisen kautta. Korkeakoulupedagogiikan ajan henki taas haastaa kehittämään opetusta yhdenvertaisuutta ja inklusiivisuutta painottaen. Tampereella ruotsin pakollisen opintojaksomme arviointi perustuu nykyisellään lopputentteihin – syyt ovat pitkälti käytännöllisiä, sillä jatkuvan arvioinnin oppimista tukevista eduista opitaan jatkuvasti lisää (esim. Liontou 2023; Mäkipää ym. 2023). Pedagogisia näkökulmia kirjallisen kielitaidon arviointiin on lukuisia ja haluaisimme tuoda ne keskusteluun erityisesti työelämälähtöisyyden ja tasa-arvoisuuden taustoja vasten.
Kutsumme kollegat mukaan kielikeskusten väliseen keskustelevaan työpajaan Kirjallisen kielitaidon arvioinnin nykyisyys ja tulevaisuus seuraavilla teemoilla:
- Kielimallit, apuvälineet ja arviointi (pj. Miro Metsämuuronen)
- Kielitaito ja kirjallisen osaamisen arviointi (pj. Henriikka Rinta)
- Inklusiivinen arviointi (pj. Susanna-Hakulinen-Bernard)
Työpaja alkaa alustuksella (10 min), minkä jälkeen jakaudumme keskustelemaan teemaryhmiin (25 min). Ovatko inklusiivinen arviointi ja apuvälineiden käyttö kielten opetuksessa tulevaisuutta? Mitkä ovat haasteet? Miten niihin vastataan?
Aihetyöpajojen havainnot kerätään ylös ja vedämme ne yhteen lopuksi. Toiveena on myös herätellä keskustelua tulevaisuuden yhteistyöstä.
Lähteet
Liontou, Magdalini. (2023). How do Finnish and Chinese students’ diverse pedagogical experiences shape feedback interpretation? AFinLA-teema 15, 81–96.
Mäkipää, Toni; Raili Hilden & Ari Huhta (2023). Oppimista tukeva arviointi kieltenopetuksessa. AFinLA-teema 15, 5–18.
Nieminen, Juuso Henrik (2021). Governing ‘the disabled assessee’: a critical reframing of assessment accommodations as sociocultural practices. Disability & Society, 37:8, 1293-1320.
Julie Uusinarkaus, Lauri Mäkela, Edward Trethowan and Leena Evesti, University of Helsinki and William Hellberg, Aalto University
ChatGPT: A Practical Tool? Evaluation and Discussion
Language professionals have regarded Chat GPT and other AI-powered large language models with suspicion. Since our job is to produce good-quality text, we wonder whether they will start taking over part of our role. But is this the only direction for the future, or is there a way that we can take these systems into use to make our roles more effective? In the end, we hope to turn what might be a threat into a promise.
In this workshop, we will explore the potential of Chat GPT and take a look at its competitors, PaLM from Google and Azure from Microsoft. Our approach is a practical one: start from small, everyday tasks to understand how we can use the model interactively and proactively. Instead of being forced to adapt to new forms of technology competing with us, how can we adapt Chat GPT for use on our own terms?
In the roundtable, first we will review what ChatGPT is and how it works, and present PaLM and Azure and a brief history of each. Then each of the roundtable presenters will share their own trial of ChatGPT for a task that is common in their own field of work. We will cover translation, editing and revision, and administrative tasks in language services. Each case study will be evaluated for how ChatGPT was a help, or if the task was a failure in the end. Then we will talk through any questions people have, and if time permits, do some hands-on work with the model.