The comprehensive dichotomy of big data evidence:An analysis based on the dichotomy between physical evidence and testimonial evidence
The massive volume of big data cannot be fully inspected by humans,but can only be converted by algorithms to big data report,which can be used as evidence.Depending on whether the algorithm incorporates information from outside the big data set,big data report can be comprehensively divided into two categories:analytical report and comprehensive report.This classification aligns with the dichotomy between physical evidence and testimonial evidence.Physical evidence is characterized by providing direct perception of facts.As long as cognitive tools accurately reflect the evidential information and do not cause evidential information's impairment within the scope of the purpose of proof,the directness of such perception will not be affected.Analytical big data report provides direct perception and thus is essentially physical evidence.On the other hand,testimonial evidence provides indirect perception of facts.The factfinder cannot directly perceive facts by their own perceptions,but can only relies on witnesses'and experts'statements to form indirect perception of facts.Comprehensive big data report provides indirect perception and is essentially testimonial evidence.The conformity of analytical big data report,which is deemed as physical evidence,with the big data set can be verified directly by the factfinder.However,the assessment of the conformity of comprehensive big data report,which is deemed as testimonial evidence,with the big data set cannot be separated from the interpretation of the algorithm.Since the algorithm's self-explanation is unreliable,such interpretation should be conducted by experts.
Big data evidencePhysical evidenceTestimonial evidenceAlgorithmGenerative artificial intelligence
NETL
NSTL
万方数据
