Abstract:In order to improve the security protection capability of big model, this paper studies the improvement path of enterprise data security capability in the era of big model. Through the systematic analysis of the security risks generated by the large model technology, compared with the security framework and security standards of the mainstream large model manufacturers, the combination security framework design and practical testing based on the security model and the model security barrier are proposed. The results show that this study can provide theoretical reference and practical path for enterprises to cope with the challenges of network security in the AI era.