Brain damage in newborns is difficult to detect. Cerebral palsy (CP) is often not detected until 1.5-2 years of age. However, an earlier intervention can reduce lasting functional impairments in the child. Karolinska Institutet (KI) has allowed parents to record approx. 1000 films on 400 children. Linköping university (LiU) will now prototype an AI system that can assess risk for CP based on the child's body and hand movements. The are several benefits to an AI approach: - Prematurely born newborns can avoid permanent damage - The parents become involved. An early diagnosis reduces their anxiety and stress - Specialist care is relieved - Societal gain: Diagnosis can be given regardless of distance to specialist care Today, the children are assessed according to the General Movement Assessment (GMA) method, but this is time-consuming and requires specialist doctors with special training in GMA. We want to develop machine learning to perform an initial screening for abnormal movement patterns in videos uploaded by parents. The technology will relieve specialist care, and will enable earlier detection of CP. The project has been granted ethical approval: 2018/236-31