Home | News

24.09.2026

Teaser image to Barbara Plank: Building Language Technology for Everyone

Barbara Plank: Building Language Technology for Everyone

Research Film

Language is full of variation. So why are AI systems built as if it isn’t?

«German is considered one of the well-resourced languages that AI can understand very well. But there is a lot of variation within German that's unexplored.»


Barbara Plank

MCML PI

As a dialect speaker herself, MCML PI Barbara Plank has firsthand experience with language technologies that don’t support her own dialect. This has motivated her research to develop Natural Language Processing (NLP) systems that are accessible to everyone, regardless of how well-supported their language or dialect is. She describes the challenges posed by language ambiguity for NLP systems and the risk of developing monotone AI systems that threaten language variation.

In this video series, discover how MCML researchers are bridging theory and practice to build AI systems that prioritize accessibility, sovereignty, and impact.

The film was produced and edited by Nicole Huminski and Nikolai Huber.

#blog #research #plank

Related

Link to Center for Digital Medicine and Health expands

16.09.2026

Center for Digital Medicine and Health Expands

TUM's Center for Digital Medicine and Health hits a construction milestone, advancing AI and medicine research.

Read more
Link to Vincent Fortuin Receives ERC Starting Grant for Reliable Small-Data AI

16.09.2026

Vincent Fortuin Receives ERC Starting Grant for Reliable Small-Data AI

MCML PI Vincent Fortuin receives an ERC Starting Grant for AutoBayes, advancing reliable AI methods for learning from small datasets.

Read more
Link to Björn Eskofier: The Missing Ingredient for AI in Medicine Is Real-World Data

14.09.2026

Björn Eskofier: The Missing Ingredient for AI in Medicine Is Real-World Data

MCML PI Björn Eskofier explains why real-world data is key to using AI for personalized treatment and more empowered healthcare.

Read more
Tiny logo
Link to MCML at ECCV 2026

07.09.2026

MCML at ECCV 2026

MCML researchers are represented with 28 papers at ECCV 2026.

Read more
Tiny logo
Link to MCML at ECML-PKDD 2026

04.09.2026

MCML at ECML-PKDD 2026

MCML researchers are represented with 6 papers at ECML-PKDD 2026.

Read more
Back to Top