Abstract--The size of deep learning models in artificial intelligence (AI) software is increasing rapidly, hindering the large-scale deployment on resource-restricted devices (e.g., smartphones). To mitigate this issue, AI software compression plays a crucial role, which aims to compress model size while keeping high performance. However, the intrinsic defects in a big model may be inherited by the compressed one. Such defects may be easily leveraged by adversaries, since a compressed model is usually deployed in a large number of devices without adequate protection. In this article, we aim to address the safe model compression problem from the perspective of safety-performance co-optimization. Specifically, inspired by the test-driven development (TDD) paradigm in software engineering, we propose a test-driven sparse training framework called SafeCompress. Then, considering two kinds of representative and heterogeneous attack mechanisms, i.e., black-box membership inference attack and white-box membership inference attack, we develop two concrete instances called BMIA-SafeCompress and WMIA-SafeCompress. Further, we implement another instance called MMIA-SafeCompress by extending SafeCompress to defend against the occasion when adversaries conduct black-box and white-box membership inference attacks simultaneously. We conduct extensive experiments on five datasets for both computer vision and natural language processing tasks. The results show the effectiveness and generalizability of our framework. We also discuss how to adapt SafeCompress to other attacks besides membership inference attack, demonstrating the flexibility of SafeCompress. Currently, AI software, with DNN as representatives, Model compression aims to compress a big DNN model is recognized as an emerging type of software artifact to a smaller one given specific requirements, e.g., parameter (sometimes known as "software 2.0" [2]). Rashly of DNN-based AI software has increased rapidly in recent compressing a model may lead to severe degeneration in the years (mostly because of a trained deep neural network AI software's task performance such as classification accuracy. For instance, a state-of-the-art model of computer To balance memory storage and task performance, many compression vision contains more than 15 billion parameters [3]. A recent approaches have been proposed and deployed [7], natural language model, GPT-3, is even bigger, surpassing [8]. For example, Han et al. [8] prune AlexNet [1] and reduce 175 billion parameters; this situation requires nearly 1TB of its size by 9 times while losing only 0.01% accuracy in image space to store only the model [4].