人工智慧於司法實務之可能運用與挑戰

人工智慧乃是以電腦程式模仿人類思考模式之結果，以更有效率之電腦運算，進行資料蒐集、分析。現行的人工智慧在司法審判中可用作法律檢索分析、風險評估工具、計算量刑基準、鑑識輔助系統，在專利商標實務中作為判斷侵權及核駁之依據，在執法上也可用於推算出可能的犯罪熱點而進行預測性執法。人工智慧雖然可有效運用在司法實務上，但其可能無法處理法律邏輯與常識之問題，可能對特定族群產生歧視或是對於隱私權產生侵害。另系統本身之透明性與可受監督上亦存有疑問，且在實際應用時可能受制於法令上的限制。準此，本文就人工智慧於司法實務上之可能運用與挑戰，進行全面性地探討，並提出結論與建議。

關鍵字

人工智慧；風險評估工具；量刑基準；正當法律程序；鑑識科學；預測性執法；歧視；隱私權；合理懷疑；一般資料保護規則；透明與監督

並列摘要

Artificial Intelligence (＂AI＂) is a computer program imitating human thinking process, which uses efficient means to gather and analyze data. The AI technology has already been used in judicial practice for legal analysis, risk assessment, sentencing guideline, and forensic science. This technology could also be used to determine patent and trademark infringement and prosecution. Moreover, AI can also be used in analyzing crime hotspot and subsequently in predictive policing. Despite these benefits, some challenges indeed exist when applying AI in judicial practice. Artificial Intelligence could hardly deal with issues relating to logic and general knowledge. The AI decision-making process often creates discrimination and privacy concerns. Additionally, it is worth debating how to make the algorithm transparent and supervised. Using the algorithm and personal data should also comply with current legal rules. This article aims to address all the possible applications and challenges while using AI in judiciary, and then to conclude and suggest for the future usage.

並列關鍵字

Artificial Intelligence (AI) ； Risk-Need Assessment Tool ； Sentencing Guideline ； Due Process ； Forensic Science ； Predictive Policing ； Discrimination ； Privacy ； Reasonable Suspicion ； General Data Protection Regulation (GDPR) ； Transparency and Supervision