Scholarlink

<p>Artificial intelligence (AI) — spanning predictive risk models, large language models, algorithmic decision systems, and digital-care devices — is being deployed with increasing ambition across social welfare systems worldwide. Existing scholarship documents many resulting harms — bias, opacity, surveillance, erosion of discretion — but remains conceptually fragmented and rarely anchored in a framework specific to social work. This paper develops a tri-lens analytical matrix crossing three moral traditions (utilitarian, deontological, virtue-ethical) with AI’s two operational arenas (frontstage client-facing systems and backstage algorithmic administration) and four constitutive social-work values (privacy, self-determination, social justice, human dignity). Rather than assuming AI integration is either inevitable or uniformly beneficial, the framework asks: under what conditions — if any — can specific forms of AI function as augmented intelligence that extends relational practice, rather than as automated welfare that entrenches exclusion and erodes moral responsibility? The matrix is applied to three purposively selected cases — the Allegheny Family Screening Tool (US), algorithmic welfare administration (Dutch SyRI, Australian Robodebt, and Korean crisis-household detection), and digital-care technologies (Korea and OECD comparators) — whose cross-modal heterogeneity demonstrates that distinct AI modalities pose distinct ethical risks requiring distinct safeguards. The analysis shows that AI is ethically defensible in social work only when it augments practitioner judgment without displacing relational authority, advances substantive rather than merely formal equity, and is embedded in contestable, auditable institutions. The paper’s contribution is threefold: it reframes the field’s debate from replacement versus resistance to augmentation versus automation; supplies a reusable diagnostic matrix; and integrates Korean, European, and North American evidence within a single analytic lens anchored in social work’s commitments to relational practice, care ethics, and anti-oppressive praxis.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

An ethical framework for assessing artificial intelligence as augmentation or automation in social work

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年8月5日

查看全文
  • Kwangseon Hwang

摘要

Artificial intelligence (AI) — spanning predictive risk models, large language models, algorithmic decision systems, and digital-care devices — is being deployed with increasing ambition across social welfare systems worldwide. Existing scholarship documents many resulting harms — bias, opacity, surveillance, erosion of discretion — but remains conceptually fragmented and rarely anchored in a framework specific to social work. This paper develops a tri-lens analytical matrix crossing three moral traditions (utilitarian, deontological, virtue-ethical) with AI’s two operational arenas (frontstage client-facing systems and backstage algorithmic administration) and four constitutive social-work values (privacy, self-determination, social justice, human dignity). Rather than assuming AI integration is either inevitable or uniformly beneficial, the framework asks: under what conditions — if any — can specific forms of AI function as augmented intelligence that extends relational practice, rather than as automated welfare that entrenches exclusion and erodes moral responsibility? The matrix is applied to three purposively selected cases — the Allegheny Family Screening Tool (US), algorithmic welfare administration (Dutch SyRI, Australian Robodebt, and Korean crisis-household detection), and digital-care technologies (Korea and OECD comparators) — whose cross-modal heterogeneity demonstrates that distinct AI modalities pose distinct ethical risks requiring distinct safeguards. The analysis shows that AI is ethically defensible in social work only when it augments practitioner judgment without displacing relational authority, advances substantive rather than merely formal equity, and is embedded in contestable, auditable institutions. The paper’s contribution is threefold: it reframes the field’s debate from replacement versus resistance to augmentation versus automation; supplies a reusable diagnostic matrix; and integrates Korean, European, and North American evidence within a single analytic lens anchored in social work’s commitments to relational practice, care ethics, and anti-oppressive praxis.