Here, we continuously compile all our research on AI. Don’t hesitate to reach out if you have questions or are interested in collaborating.
Unequal Use of AI
AI chatbots like ChatGPT have sparked lively discussions about their role in teaching and learning within higher education. This study examines how students are adopting and perceiving ChatGPT and similar tools. Drawing on survey responses from nearly 6,000 students across Swedish universities, the research uses descriptive and inferential statistics to explore usage patterns, attitudes, and concerns, as well as how these vary by gender, academic level, and field of study.
The results show that ChatGPT is widely recognized and used among students, while other AI chatbots remain less familiar. Over half of the students expressed positive views about using chatbots in education, though nearly as many voiced concerns about their future implications. Significant differences emerged between groups, particularly by gender and field of study. Female students and those in the humanities and medicine were more likely to express skepticism and concern about AI’s role in learning and assessment. In contrast, male students and those studying technology and engineering reported higher usage rates and greater optimism.
These findings highlight the ongoing importance of considering student backgrounds in the adoption of new technologies. They also underscore the need to address challenges related to AI’s integration into education. The study emphasizes the value of developing tailored solutions that align with students' attributes and needs, offering valuable insights for developers, educators, and policymakers.
Project uses AI to revolutionize qualitative data analysis
A groundbreaking interdisciplinary project at Chalmers is developing innovative methods using AI. The goals are to make the traditionally time-consuming process of qualitative data analysis more feasible and reliable while enabling multilingual researchers to work with their data in its original language.
Current tools are primarily designed for English, creating challenges for researchers working with data in other languages, especially in multilingual teams. This project aims to overcome these barriers by developing AI-driven solutions that allow direct analysis in multiple languages. These solutions save time and resources while preserving critical cultural and contextual nuances often lost in translation.
The project is led by Baraa Khuder, Senior Lecturer, and Raffaella Negretti, Professor, both from the Department of Communication and Learning in Science, alongside Nora Speicher, Research Specialist at Chalmers E-commons.
By simplifying the analysis of multilingual data and reducing translation bias, the project enhances the accessibility of qualitative research methods across disciplines. Researchers, particularly in STEM fields, often avoid qualitative methods due to the time-intensive processes and translation challenges when working in global teams.
“Our solution enables researchers to extract insights directly from their data while preserving both integrity and context, without extensive translations,” explains Raffaella Negretti.
The project utilizes AI’s capabilities to recognize patterns and organize data, significantly reducing analysis time and allowing for the handling of larger datasets with high precision. A key feature of the framework is its ability to preserve linguistic and cultural nuances in data in the original language—something traditional translation-based methods often fail to achieve.
The study also highlights where AI is most beneficial, such as automating routine tasks, and where human insight is essential, particularly in interpreting complex themes and contextual relationships.
The researchers believe the tool will benefit various groups in society. In academia, for instance, researchers and students will gain access to advanced AI tools while building skills in digital research methods.
By enabling faster and more accurate analysis of multilingual data, the project contributes to better decision-making in areas such as global policy, market analysis, and intercultural communication, creating benefits on a societal level. Organizations in the private sector that handle large volumes of multilingual data could also use the framework to streamline their processes while ensuring ethical AI use.
“By combining advanced technology with human expertise, the project positions Chalmers as a leader in the digitalization of research. The new framework sets a standard for how multilingual qualitative analysis can be conducted responsibly and effectively, with the potential to impact both academia and industry globally,” says Baraa Khuder.
The project will run until the end of summer 2025.
AI – Opportunity or Obstacle for Learning?
From Chat GPT – hjälp eller hinder? - Universitetsläraren (in Swedish)
How does the use of AI tools like ChatGPT influence students' learning and writing development? Raffaella Negretti, Professor of Educational Psychology and Applied Linguistics, explores how these tools can both support and hinder academic progress. With nearly 30 years of experience in teaching and researching academic writing, Negretti emphasizes the cognitive depth of writing:
“Writing is not just a skill but a way of thinking,” she explains. “By putting thoughts into words, students develop subject knowledge and critical thinking. Writing requires attention, self-regulation, and reflective thought – all essential processes for becoming a skilled writer.”
However, relying too heavily on AI tools to generate text can impede these learning processes:
“If students let AI do all the work, they miss valuable opportunities to think and develop their ideas,” says Negretti.
Despite the risks, Negretti highlights that tools like ChatGPT can promote learning when used strategically. For instance, they can help students experiment with phrasing and language, fostering their “rhetorical awareness” – the ability to adapt language and tone to different audiences and contexts.
“AI should not be used to create the first draft, as much of the learning happens during the planning and structuring of the text,” she explains. “Instead, AI can support students by offering alternatives to reflect on and discuss.”
For AI to be a support rather than a hindrance, Negretti advocates for developing critical AI literacy. This involves teaching students to:
- Reflect on the ethical and societal implications of AI usage.
- Review and evaluate the suggestions generated by AI tools.
- Use AI as a resource to enhance their own writing skills.
“We shouldn’t ban AI in classrooms. Instead, we should encourage students to use it in ways that support their learning,” she asserts.
Negretti concludes with a quote from linguist M.A.K. Halliday: “Writing is a means for thinking and a means for action.”
“ChatGPT and similar tools cannot replace the social and cognitive processes essential for becoming a writer. But when used correctly, they can be powerful tools for learning and creativity,” she emphasizes.
AI and the future of writing collaboration
Can Generative AI (GAI) tools, like ChatGPT, do more than just polish grammar? A forthcoming study explores how these tools can revolutionize academic writing by fostering deeper student engagement and enhancing their unique disciplinary voices. This structured pedagogical approach, rooted in principles of human-human interaction, reveals how thoughtful integration of GAI tools can transform the writing process.
Key Insights
- Beyond Grammar Checks: Students learned to leverage GAI tools not just for technical fixes but to spark critical thinking, ask reflective questions, and refine their ideas.
- Elevating Disciplinary Voice: By prompting targeted questions and interpreting GAI's responses, students developed a stronger and more distinct academic voice.
- Bridging Human and Machine Interaction: Strategies drawn from human-human collaboration—like contextual feedback and intentional questioning—proved effective in guiding machine-human writing dynamics.
Why This Matters:
- For Educators and Researchers: This study highlights the potential of using familiar interaction frameworks to seamlessly incorporate GAI tools into teaching and research practices.
- For Students: The findings empower students to use GAI tools critically and creatively, balancing AI assistance with independent thought.
- For Society: In a world where digital tools are integral to learning, this research advocates for responsible technology use to amplify creativity, bridge human expertise with AI capabilities, and reshape the academic landscape.
Critical AI literacy among doctoral students
The study explored how doctoral students can learn to use generative AI (GAI) effectively and ethically in their academic writing. A new concept, Critical GAI Literacy (C-GAI-L), was developed. This framework goes beyond merely knowing how to ask questions of AI tools; it emphasizes developing critical skills for interacting with AI, evaluating its ethics and limitations, and taking control of one’s own learning in this area.
Based on this concept and a strategy for self-directed learning, a micro-curriculum was designed and tested. Its goal was to help doctoral students use GAI thoughtfully and responsibly in their writing process. The results showed that students—who initially had limited knowledge and sometimes fears about GAI—developed a more nuanced understanding of the opportunities and limitations of AI. They also became more aware of issues surrounding text ownership and gained insights into how GAI affects language use and academic writing practices.
What does this mean for Doctoral-level academic writing?
The conceptual framework of C-GAI-L can support educators in designing courses and assignments that strengthen students' ability to learn independently and use GAI responsibly. For students, the framework provides practical tools to develop critical skills in applying AI to academic purposes.
By integrating Critical GAI Literacy into academic writing education, universities can prepare students to become skilled and responsible users of AI in their research fields.
Conceptualising and cultivating Critical GAI Literacy in doctoral academic writing
Urgent for universities to establish policies
This study offers pioneering empirical insights into students’ experiences with and perceptions of AI-powered language tools (AILTs) for academic communication following the introduction of ChatGPT. Surveys of over 1,700 Swedish university students revealed the widespread integration of AILTs into students’ communicative repertoires and the development of a “spatially advised learner” identity. This identity enabled students to assert agency in language learning and content knowledge while critically assessing AI’s limitations. The research also highlighted diverse student perspectives on ethical concerns surrounding AILTs in assessment in the absence of university instructions in early 2024.
Theoretically, the study advances post-humanist applied linguistics by empirically illustrating how communicative competence emerges through ‘repertoire assemblage,’ coordinating resources from human cognition and the material world. Practically, it underscores the urgency for universities to establish policies and educational support to help students navigate AI-integrated learning and assessment. Embracing a repertoire assemblage perspective can support policymakers and educators in recognizing the full spectrum of students’ communicative resources, enriching their academic experiences in the AI era.