新式測謊技術發展現況與潛在問題

The present state and potential problems of new lie detection technology

過去的測謊研究，多數仍聚焦於測謊儀上，主題包括計分方式、判讀規則、題組設計、測前晤談等。然近二十年來，因為資訊科技與設備快速發展，使得謊言偵測有了不同的面貌，除了開發許多新式設備外，也運用機器學習理論協助結果判定。由於法令與實務之需求，本文主要介紹各種非接觸式測謊設備理論與開發現況，包括熱影像、語詞、語音、臉部表情、眼動、都卜勒雷達等，部分技術已經趨近實用化，部分則因使用限制或有替代設備而未再開發。綜觀目前發展現況，新式測謊技術雖具有更多選擇與判斷方式，但存在有偵測設備、會談情境建立、取樣數量、系統整合、標作業程序與個人化分析模式建立等問題尚待克服。

"Lie detection study has been mostly concentrated on the polygraph instrument in the past. The research topic covered the scoring method, evaluation rule, test question modification, and pretest interview arrangement. Due to the rapid development of information science and computer instruments over the last decade, lie detection has taken a different route. In addition to the development of new equipment, machine learning techniques have been used to assist in the outcome determination. In order to meet the criteria of law policy and practices, this article introduces the theories and advances of several non-contact lie detection devices, such as thermal imaging, linguistic style, auditory analysis, facial expression, eye movements, and Doppler radar. Some of the technologies have progressed to the stage of practical application, while others have remained in lab due to instrument limitation or alternative approaches. Overall, contemporary lie detection technology improves selectivity and judgment processes, but there are still certain challenges in this field, such as interview scenario construction, sampling shortages, system integration, the formation of standard operating procedures and individualized analysis mode."