Many Attribute-Based Encryption (ABE) schemes do not protect receivers' privacy, such that all the attributes to describe the eligible receivers are transmitted in plaintexts. Hidden policy-based ABE schemes have been proposed to protect receivers' privacy by using a construction that requires every user in the system to decrypt the ciphertext using all the attributes they possess, which incurs great computation and communication overhead. To address this issue, in this paper, we propose a new concept Gradual Identity Exposure (GIE) to protect data receivers' identity. Our approach is to reveal the receivers' information gradually by allowing ciphertext recipients for decrypting the message using their possessed attributes oneby-one (but not all). If the receiver does not possess one attribute in this procedure, the rest of attributes are still hidden. Compared to hidden-policy based solutions, GIE provides significant performance improvement in terms of reducing both computation and communication overhead. We also present a theoretical framework to model the GIE with several new proposed concepts.