I hadn’t thought about the effects of GDPR on machine learning, but since it gives parties a “right to explanation” that seems to suggest a need for model explanations. This article talks a little about the implications of GDPR on machine learning which encompass both data management and explain-ability.

I know there are some other model explainer packages in R such as xgboostExplainer and breakDown (meant for any model). For Python, LIME seems to be the most population option although there is also Skater which I haven’t used. Hopefully there will be some growth in this space as emphasis shifts from pure accuracy to accuracy and interpretability.