Advocacy for Black and Minority Ethnic Communities: Understandings and Expectations

Recent policy has promoted ‘advocacy’ as a meansof promoting social justice for many disadvantaged groups. Yet‘advocacy’ is a contested concept, and the understandingsthat members of disadvantaged groups themselves have of advocacyhave rarely been explored. Previous research indicates thatunderstandings may vary considerably. Using empirical evidencefrom research conducted in Glasgow, Scotland, this paper examinesthe understandings and expectations of ‘advocacy’held by black and minority ethnic (BME) service providers andpotential service users. The BME service providers believedthat they were offering advocacy, and did so in the contextof a marginalized position for their services. The BME communitymembers supported the development of advocacy services, buttheir own marginalization was in many ways reinforced by servicesthey were already using. They had clear ideas about appropriateadvocacy services for their situation. These ideas were groundedin their current situation, and did not necessarily conformto dominant ideas about advocacy. In conclusion, the prospectsfor successful advocacy are assessed.     

人工智能介入量刑机制：困境、定位与解构

司法人工智能分为常规型人工智能与专业型人工智能,前者是将通用领域已经发展成熟的人工智能直接移植至司法领域而无需专门的算法更新,主要目的是将审判人员从繁重的"事务性"工作中解放出来,因而其也无法介入审判的核心内容;后者是诸如量刑辅助系统等专门为司法领域开发的介入审判实质内容的人工智能,其是司法人工智能的核心。当前司法人工智能的实践现状是常规型人工智能因其有坚实基础而卓有成效,但极为重要的专业型人工智能的开发与使用并不理想,主要原因是法学研究对专业型人工智能的研发理论供给不足,具体表现为"抽象有余而具象不足",其深层次原因在于法学专业知识与人工智能技术知识没有深度融合,即"懂技术的不懂法律,懂法律的不懂技术"。宏观层面,在智能爆炸不可预期的时空背景下,生命2.0阶段(文化阶段)或弱人工智能时代仍是当下及可预见未来所长期处于的阶段,故作为"工具"的量刑人工智能仍应定位于辅助量刑而非决定量刑,且基于量刑规范化改革的价值内涵,应更进一步地定位于规范性辅助而非参考性辅助,二者的区别是智能系统给出的阶段性量刑结论对法官的约束力大小。在微观层面,智能量刑系统的算法构建应以量刑逻辑主导算法逻辑为原则,以诸如量刑基准、不法刑等具有"共性"属性的阶段性量刑为作用领域而非其能力之外的终局性量刑结论(宣告刑);此外,为防止算法黑箱、算法歧视以及相关关系代替因果关系,须做到量刑人工智能的算法公开和阶段性量刑结论的可解释性。