Research on the Authentication of Big Data Evidence

Abstract

As one emerging evidence, big data evidence is mostly presented in the form of electronic data. The authentication of evidence based on data technologies requires strengthened examination of the authenticity of the original electronic data, the identity and integrity of the objective electronic data, and the reliability of the presented electronic data. In judicial practice, authentication tends to be defined within the physical ambit regarding the carriers and information, with the emphasis on the “chain of custody” and “uniqueness identification”; however, the technical authentication is still underdeveloped, such as the integrity check value, read-only with mirror copies, and intelligent authentication. Besides, the examination of the reliability of algorithm is overly neglected, which includes the assessment of “the training dataset”, accuracy, adaptability, interpretability and reversibility, and further weaken the trial and make the materialized examination of big data evidence impossible. Therefore, it is necessary to clarify the concept of big data evidence, and then to define the object of authentication and construct corresponding authentication rules, which focuses on the comprehensive authentication both on the physical and informational level, thus responding more scientifically to practical needs.