Deep Case-AI based Crime Investigation System

Research Goal

The Korean National Police Agency announced the termination of police investigation, which has lead urgent need to minimize error rates for final investigation results. The necessity of reviewing errors in investigations and legal issues has been raised, mainly by investigators with lawyer qualifications. However, individual reviews of more than 1.7 million cases are said to have limitations.

The Korean National Police Agency has promoted to utilize AI (Aritifial Intelligence) technology for criminal investigation to reduce the amount of time and decrease legal loopholes during the long-term investigation period of each case. The main goal of the “Deep Case-AI based Crime Investigation System” is to develop an automated AI-based Crime Investigation System using NLP and argument mining.

Beyond data analyzation, tasks such as similar case matching, legal inference reasoning, and developing visual argument structure algorithm tasks etc. are conducted. The exclusive means of this project is to enhance quality, thus develop an AI-based criminal investigation technology that quickly reviews issues of a case based on the systematic and analytical algorithm that can be generally applied to all cases.

Key Technology

The five core technology are :

Through these key technologies, we can verify and interpret objective criminal investigation verfification real-time, recommend similar events and visually compare arguments, rank leading alternative hypothesis and supplementary requirements based on guidelines to investigate legal documents.

Expected Outcomes

  1. Technical aspect : Secure AI technology applicable to the criminal investigation field.
  2. Economic aspect : Drastically shorten case analysis time & save labor costings.
  3. Social aspect : Enhance reliability of the police investigations by ensuring objectivity.

Video of ‘Deep Case-AI based Crime Investigation System’ project :

Deep Case-AI based Crime Investigation System