巨量資料發布系統之個人資料去識別化要求事項初論

Abstract

隨著「大數據」、「資料探勘」的興盛發展，保護個人隱私的「去識別化」相關技術與標準化的需求也日益受到重視。相關的隱私洩漏事件已清楚證明了資料庫的公開往往隱藏著當事人的資料被識別的風險；為此，各方提出了各種保護個人隱私資料去識別化的實作方法與訂定相關標準。自2014年11月17日法務部法律字第10303513040號函之函釋：「去識別化之個人資料依其呈現方式已無從直接或間接識別該特定個人者即非屬個人資料」起，如何制定「去識別化(De-identification)」要求事項已成焦點。為了平衡個人隱私保護之風險與公開資料分析的效益，本文闡明前述去識別化之內涵，並建議在我國「個人資料去識別化過程驗證要求及控制措施」的驗證規範之要求事項與控制措施中宜考慮加入相關規範。

With the progress of “big data” and “data mining”, “de-identification” techniques and standards that protect people’s privacy had become more and more important. Privacy leak cases had proved the risk that people’s data being identified come hand in hand with database opening. To solve the privacy protection problem, different parties have come up with a variety of de-identification methods and standards. Since 2014, Nov 17th the Ministry of Justice has explained that “If De-identified personal information cannot identify directly or in-directly a specified individual, it is not personal information.” How to Formulate “De-identification” requirements has become a hot research topic. In order to achieve the balance between the risk of personal privacy violation and the benefit of open data, we formulate the standardization content of de-identification and make suggestions also the additional requirements the government and take to protect personal privacy in data base opening.