Accessible communication: AI-assisted translation into plain language

'T2K' stands for ‘Text2Knowledge’ – and the name says it all: together with Dresden University of Applied Sciences, this start-up is seeking to make the automatic translation of German text into plain language possible.

Two men and one woman smiling into the camera. — The team of the project idea 'Accessible communication: AI-assisted translation into plain language'

Why are you a strong team?

The team behind ‘Accessible communication: AI-assisted translation into plain language’ comes from project partners t2k GmbH, HTW Dresden and Johannes Gutenberg University of Mainz, combining different subject areas, genders and backgrounds. The link between artificial intelligence, computational linguistics and translational studies is a strong expertise combination and an outstanding foundation for solving the problem of plain language.

Explain your idea in three sentences.

Societal and technological challenges are steadily increasing. Expressing them in words is difficult, as the increased complexity associated with them is also reflected in language.

The ‘Accessible communication: AI-assisted translation into plain language’ project addresses a fundamental, societal problem, with language and communication being simplified through AI and also made accessible for those who cannot read and understand difficult texts.

What makes your idea special?

Complex language excludes many people from information flows. On the other hand, the German Disability Equality Act, BITV 2.0 regulation and debate surrounding an official standard for plain language underscore the need for broad information accessibility. Our project will enable German text to be simplified quickly and automatically, making a significant contribution to social participation by ALL.

What are the next steps?

Existing corpora for language simplification and plain language will be aggregated and expanded. Furthermore, a concept for measuring and assessing language simplification will be developed and user-group-specific requirements identified (‘who needs simplification and to what degree do they need it?’). We will train an initial model for automated language simplification.