提示詞中無意洩露機密資訊

Risk Description

Users unintentionally input confidential business information, personal privacy data, or protected intellectual property in prompts during AI interactions. This information may be logged by AI service providers, used for model training, or indirectly accessible through other users queries in multi-tenant environments. Enterprise employees using public AI services to process internal documents has become a common source of major security incidents.