In the era of Internet of Things (IoT) technologiesthe potential for privacy invasion is becoming a major concernespecially in regards to healthcare data and Ambient AssistedLiving (AAL) environments. Systems that offer AAL technologiesmake extensive use of personal data in order to provide servicesthat are context-aware and personalized. This makes privacypreservation a very important issue especially since the usersare not always aware of the privacy risks they could face.A lot of progress has been made in the deep learning field,however, there has been lack of research on privacy preservationof sensitive personal data with the use of deep learning. In thispaper we focus on a Long Short Term Memory (LSTM) Encoder-Decoder, which is a principal component of deep learning, andpropose a new encoding technique that allows the creation ofdifferent AAL data views, depending on the access level of the enduser and the information they require access to. The efficiencyand effectiveness of the proposed method are demonstratedwith experiments on a simulated AAL dataset. Qualitatively, weshow that the proposed model learns privacy operations such asdisclosure, deletion and generalization and can perform encodingand decoding of the data with almost perfect recovery.