預測性警察活動在犯罪偵防運用與問題

The Problems and Implementation of Predictive Policing on Criminal Justice

近年來隨著人工智慧（AI）與大數據（big data）的發 展，警察活動逐漸走向數據導向與預測性，逐漸形成預測性 警察活動（predictive policing）的風潮。這項發展在美國結 合到電腦軟體，開發出多樣以地點、人與時間的預測性警察 活動程式，運用到犯罪偵防與偵查上，雖然成效卓著，但同 時也引發諸多問題。 本文的主要目的即在探討此問題，從預測性警察活動的 演進、運用，乃至於所衍生問題，做一個鳥瞰式的評估，並 從美國學者的論述，檢討我國未來可能發生問題，找出一條 解決之道。

Accompanying to development of AI and Big data, the police acts as well as data-driven and predictive. It trends to Predictive policing. In American, it combines with computer software and develops many programs. There are many sorts of place, target and time. It brings effectiveness and problems as the same time. The main purpose of this article is trying to explore these problems. It describes the evolution, implement and the problems of predictive policing. And it gives a evaluation about predictive policing from the American scholar's literatures. It try to find a way to solution the future problems when predictive policing occurs in Taiwan.