법과학 증거의 정량화에 관한 법적 연구

• 영문명: A Legal Study on Quantified Forensic Evidence: Focused on Probability of Evidence and the Possibility of Cognitive Bias
• 발행기관: 한국형사정책학회
• 저자명: 이성기(LEE, SEONGKI) 조영일(CHO, YOUNGIL)
• 간행물 정보: 『형사정책』刑事政策 第28卷 第3號, 185~210쪽, 전체 26쪽
• 주제분류: 법학 > 법학
• 파일형태: PDF
• 발행일자: 2016.12.30

영문 초록

This article discusses current a forensic evidential issue on how to quantify forensic evidence. As it has been remarkedly noted in the report of National Academy of Sciences(NAS), pattern evidence, except for DNA evidence, has shown no scientific basis and as a result this type of evidence exposes fact finders to cognitive bias making them accept the evidence as individualized identification. For this reason, the NAS report recommends that forensic science institutes identify forensic evidence by using a quantified method such as likelihood ratio(LR), random match probability(RMP), and verbal equivalents(VE). According to current psychological research, however, there is a risk of cognitive bias when lay people interpret those quantified forensic evidence such as prosecutor’s and defense attorney’s fallacy requiring that fact-finders in trial court to be informed such risks to avoid any possible error in a decision process. In this sense, it is a paradox to scientists and legal professionals that to the former quantified forensic evidence is scientific and logical whereas it is only conditionally acceptable to the latter. In Korean setting, however, it is problematic that there is not ample of meaningful discussion on quantified forensic evidence among legal professionals. This article suggests that Korean forensic science institutes start by quantifying fingerprints because the institutes store almost all Korean fingerprint data. And this paper also recommends further persistent academic dialogue between legal and forensic disciplines.

목차

Ⅰ. 들어가며

Ⅱ. 법과학 증거의 정량화와 관련된 외국의 변화

Ⅲ. 정량화된 법과학증거의 인지적, 법적 문제

Ⅳ. 개선방안을 위한 몇 가지 제언

Ⅴ. 맺으며

키워드

법과학 증거 지문 증명력 개연성 인공지능 Forensic Evidece Finger print Weight of Evidence Probability Artificial Intelligence