Johannes Bjerva

Professor · Natural Language Processing

Linguistically grounded NLP for trustworthy AI.

I am Professor of Natural Language Processing at Aalborg University and lead AAU-NLP. My fundamental research investigates how language-based AI systems encode and use meaning, what shapes generalisation across languages and cultures, and when and why they fail. These insights inform work on trustworthy language technologies, with particular emphasis on security, privacy, and factual reliability—from multilingual models to increasingly complex AI systems.

Explore my research Selected publications
Johannes Bjerva
Aalborg University · Copenhagen

At a glance

Single-PI external awards

≈ DKK 25M

Plus ≈DKK 10M in shared or internal programmes.

Funding portfolio

Selected publication outlets

  • ACL
  • EMNLP
  • TACL
  • Computational Linguistics
  • AAAI
  • ICML

Peer-reviewed work across NLP, linguistics, AI security, and machine learning.

Selected publications

Selected highlights

Recent work

A selective timeline of research and career milestones.

  1. Three papers at EMNLP 2026

    Three quite different papers were accepted: on language similarity in NLP, information structure in vision-language models, and semantic leakage from image embeddings. This is especially meaningful to me as it's a special milestone for the people behind them - these are the final PhD thesis papers for Yiyi and Marcell, and the first for Tao.

  2. Typologically diverse evaluation

    Published in Computational Linguistics, bringing several years of work on defensible multilingual evaluation into one framework.

  3. Coefficient Giving · Formal semantics and AI safety

    Opened a new DKK 2.4M programme connecting formal semantics with multilingual AI-safety analysis.

  4. DFF Sapere Aude · TRUST

    Made it possible to establish a sustained programme in linguistically grounded language-model security.

  5. NLP Security and Ethics, in the Wild

    A TACL paper setting out a research agenda for security and ethics in real-world NLP systems.

  6. Full professor at Aalborg University

    Appointed full professor of Natural Language Processing, opening a new phase for NLP research at AAU Copenhagen.

  7. Novo Nordisk Foundation · (LM)²-SEC

    Gave a growing team five years to pursue security and privacy in multilingual language models.

  8. Carlsberg Foundation · Semper Ardens: Accelerate

    A formative DKK 5M award for multilingual modelling and the growth of the NLP environment at AAU.

Collaboration

For research, media, or policy collaboration.

Get in touch