Prosecutorial reform and the creation of an independent charging agency open the door for AI-assisted indictment systems, gradually replacing human discretion with algorithmic recommendations in criminal justice.
Following sweeping prosecutorial reform, South Korea creates a Public Charging Agency separated from the existing prosecutor's office. Facing staffing shortages and political pressure to demonstrate objectivity, the agency adopts an AI system trained on two decades of case outcomes to recommend charging decisions. Initially positioned as advisory, the system proves statistically more consistent than human prosecutors in predicting conviction rates. When media reports reveal that human overrides correlate with political cases, public trust shifts decisively toward the algorithm. The agency introduces mandatory justification panels for prosecutors who deviate from AI recommendations. Defense attorneys discover that the system systematically under-weights circumstantial evidence and over-indexes on prior records, effectively criminalizing poverty and repeat offenders while giving first-time white-collar criminals favorable assessments.
Prosecutor Lee Seokhwan stares at the red 'DECLINE CHARGE' recommendation on his screen for a domestic violence case he knows in his gut should go to trial. The victim recanted under pressure — the algorithm reads that as weak evidence. He has forty minutes to write a formal override justification that will be reviewed by a five-member panel, three of whom have never tried a case. He opens the form, then closes it. He opens it again. The cursor blinks on an empty page.
Algorithmic consistency could eliminate the most egregious forms of prosecutorial abuse — politically motivated investigations, personal vendettas, and unconscious bias in charging decisions. A transparent, auditable system might actually deliver more equitable justice than a system dependent on the mood, workload, and political ambitions of individual prosecutors.