AI LAB for Low-Literacy

AI LAB for Low-Literacy

Low-literacy is a substantial problem in the Netherlands. Approximately 2,5 million adults were functionally illiterate in 2016 (Algemene Rekenkamer), and their number will rise: recent (inter)national assessments have shown a steady decline in the (deep) reading comprehension skills of Dutch teenagers from 2012 onwards. Dutch children learn to read and write in primary school: according to their final tests (often the CITO-test), they perform adequately at the age of 12. However, according to the international PISA survey, a quarter of them are at risk of developing low-literacy at the age of 15. What factors contribute to the drop in literacy between the age of 12-15 and how can this drop be prevented? Identifying these factors has proven to be too complex for individual researchers and their data sets: it requires AI technology that is central to our AI LAB for Low-Literacy.

The AI LAB for Low-Literacy aims to design tools that model the factors crucial for the risk of developing low-literacy between the ages 12 and 15 at the individual and population level. At the individual level we aim to develop AI technology that specifically models problems with literacy by means of speech and language processing: Oral assessments provide the possibility of a fine-grained analysis of individual reader’s developments. At the population we aim to address at-risk schools and pupils and to develop and test a personalized AI support tool in the defined target population (from first to third year high school students) to intervene and guide pupils to appropriate reading and writing course modules.

As a proof of concept, we particularly develop personalized AI guidance and assistance to support personal autonomy and intrinsic motivation of the at-risk target population.

The core team of the AI LAB for Low-Literacy consists of:

Hans Marien (Lab Coordinator; Social and Behavioural Sciences)
Els Stronks (research advisor: literacy and language processing)
Henk Aarts (research advisor: social cognition and human behavior)
Mehdi Dastani (research advisor: Agent-based AI systems) behavior)

Currently, the AI LAB includes collaboration with the Foundation for Open Speech Technology, Fontys University of Applied Sciences, Royal Library, Stichting Lezen.

Website: coming soon!