We are very happy to have been selected for the first Challenge for Europeana AI/ML Datasets by EuropeanaTech, along with two other projects.

We proposed to build ConConCor, a corpus of contentious terms in context based on a selected subset of the Dutch newspaper collection within Europeana. With the provided funding, we will crowd-source annotations of terms, their level of contentiousness and preferred alternatives. Once ConConCor is established, we will conduct a pilot study into how it can be used to bootstrap and evaluate bias-aware AI.

The project will be completed by the end of June, we're excited to get to work!