Language & meaning
How do models handle differences between languages? We study meaning and multilingual representations, including languages with limited training data.
Explore language & meaningNatural Language Processing · Aalborg University
Language is where AI meets the world.
We research trustworthy language AI: how models represent meaning and knowledge, and how to make their use secure and factually reliable. Our work connects computational linguistics with multilingual NLP and knowledge representation.
5faculty
18current researchers across career stages
9current and forthcoming projects
2interdisciplinary labs
Research areas
What we study
We investigate how models handle language and knowledge, focusing particularly on security and factuality. We study how to protect models from attacks and privacy leakage, and how to ground their outputs in evidence.
Knowledge & reasoning
How can models use structured knowledge? We study graph representation and reasoning, alongside the evidence behind generated claims.
Explore knowledge & reasoningSecurity & reliability
Where do models fail or expose private information? We test multilingual vulnerabilities and develop methods for reliable evaluation.
Explore security & reliabilityLanguage technology in use
Whose needs do language tools serve? We study community priorities and applications, as well as AI’s effects on information diversity and the environment.
Explore language technology in useFaculty
Meet our faculty
Our faculty work on different questions and collaborate where their interests meet.
Johannes Bjerva
Professor of Natural Language Processing · Group Leader
Johannes studies how linguistic structure helps explain multilingual language models, particularly their security and privacy risks. He leads AAU NLP and AAU's Copenhagen Section of Computer Science.
Russa Biswas
Tenure-Track Assistant Professor · Co-director, AI:PAGE-Lab
Russa studies knowledge-graph representation and reasoning, and how graphs can support factual language-model outputs. She co-directs AI:PAGE-Lab, applying these methods to historical collections.
Dustin Wright
Tenure-Track Assistant Professor
Dustin studies whether AI-generated information remains faithful to its sources and preserves diverse knowledge. He also works on efficient models and AI sustainability.
Heather Lent
Assistant Professor
Heather studies multilingual language-model security and the ethics of security research, including privacy risks from text embeddings. Her work also covers Creole and other lower-resourced languages.
Xikun Jiang
Assistant Professor · Joint affiliation with Formal Methods for Security & Privacy
Xikun develops trustworthy and responsible AI, with a focus on privacy-preserving methods and AI verification for secure and reliable systems.
Project highlights
Projects and fellowships
Two interdisciplinary labs and two forthcoming fellowships, spanning historical knowledge, AI security, information diversity and accessibility.
AI:PAGE
Interdisciplinary lab
AI:PAGE-Lab
Turns historical texts into searchable, connected knowledge so researchers can discover patterns and relationships across cultural-heritage collections.
Official recordAI:SECURITY
2025–2029
AI:SECURITY
Connects NLP and cybersecurity to protect AI systems from hijacking and misuse and to counter AI-enabled disinformation and phishing.
Official recordDARA PhD fellowship
Forthcoming · Starts 15 November 2026
Mode Collapse Robust LLMs
Investigates how alignment training narrows the range of ideas expressed by language models, and develops ways to preserve that diversity while maintaining safety.
Official recordDFF International Postdoc
Forthcoming · November 2026 – October 2028
Fast and Focused: Information-Structure-Aware AI for Accessibility
Studies how AI handles the way speakers signal familiar and new information, with the aim of improving the speed and reliability of assistive language technology.
Official recordHow we work
Setting people up for success
We share work in progress, make expectations explicit and recognise the contributions that research depends on. Our Handbook sets out how we put these commitments into practice.
Our research culture- Rigour Questions and evidence before venue or fashion.
- Feedback Share work while it can still change.
- Responsibility Independence without isolation.
- Trust Open disagreement with respect and support.
Selected work
Selected papers
Examples of our work on language and knowledge, including security research and the diversity of information generated by AI.
NLP Security and Ethics, in the Wild
Audits ethical practice in NLP security and develops concrete guidance for harm minimisation and responsible disclosure.
- Security, Privacy & Safety
- Communities, Domains & Consequences
What and Whose Knowledge? Measuring Epistemic Diversity in Large Language Models
Measures the diversity of claims in language-model outputs across topics and cultural contexts.
- Communities, Domains & Consequences
- Factuality, Reliability & Evaluation
MADLINK: Attentive multihop and entity descriptions for link prediction in knowledge graphs
Combines graph paths with entity descriptions to predict missing links in knowledge graphs.
- Knowledge, Evidence & Reasoning
Do LLMs Really Memorize Personally Identifiable Information? Revisiting PII Leakage with a Cue-Controlled Memorization Framework
Separates genuine memorisation from reconstruction enabled by lexical cues across 32 languages, tightening the evidential standard for claims about PII leakage.
- Security, Privacy & Safety
- Multilingual & Lower-Resource NLP
MultiHal: Multilingual Dataset for Knowledge-Graph Grounded Evaluation of LLM Hallucinations
Builds a multilingual multihop benchmark from 25,900 curated knowledge-graph paths for evaluating and reducing hallucinations.
- Knowledge, Evidence & Reasoning
- Multilingual & Lower-Resource NLP
When Discourse Pressures Conflict: Information Structure in Vision-Language Model Outputs
Shows that vision-language models respond to discourse context in Hungarian but compress interacting information-structural pressures into narrow patterns.
- Language, Meaning & Typology
- Factuality, Reliability & Evaluation
News & milestones
What is happening now
People
Six new researchers join AAU NLP
Since 1 June, Ren Tao, Anahita Baninajjar, Dorielle Lonke, Anna Lackner, Davis Davalos-DeLosh, and Lena Pickartz have joined the group, strengthening our work in linguistics, trustworthy AI, and language-model security.
Publications
Five papers at EMNLP 2026
Four Main Conference papers span multilingual language modelling, vision-language models, embedding security, and epistemic diversity. They include Tao's first paper and the final PhD thesis papers for Marcell and Yiyi. A fifth contribution was accepted to the System Demonstrations track.
Projects
Two new programmes begin
TRUST and Formal Semantic Methods for AI Safety strengthen the group's work on model integrity, formal semantics, multilingual failure analysis, and explainable safeguards.
Work with us
Develop a research partnership.
Building a European consortium or developing an MSCA proposal? Talk to us about LLM security and factuality, or a shared question in language technology.
Collaborate with us