독서리서치레이더투자 프레임워크
로그인 / 회원가입
← 읽기로 돌아가기
뉴스주제 · AI责任归属之争중대3 분 소요DisconfirmAI 편집부

Patrick J. McGovern Foundation President Billas Dhar: AI Harm to Consumers Demands More Than Legal Compliance, Yet Liability Allocation Remains Unresolved

Billas Dhar, president of the US-based Patrick J. McGovern Foundation, argued in a Seoul interview that corporate responsibility to consumers when deploying AI extends beyond legal compliance and must include pre-harm risk monitoring. He also identified the allocation of liability among AI developers, deployers, and third parties as an unresolved fundamental question.

2026-10-20 08:47 GMT+8

When a company uses AI to process customer orders or provide service and something goes wrong, who is responsible? Billas Dhar, president of the Patrick J. McGovern Foundation, offered a clear answer in a Seoul interview: whether a company outsources work or introduces AI systems, its responsibility to consumers does not shift. But he immediately raised a harder question—if an AI system genuinely harms a consumer, how should liability be divided among the developer who built it, the company that deployed it, and any third parties in between?

“If a corporate AI system causes harm to a consumer, is it the person who built the AI, the company that used it, or a third party who should be held accountable? That is a very big question we must answer,” Dhar said. He became the first president of the McGovern Foundation in 2020 and previously served as a member of the UN Secretary-General’s High-Level Advisory Body on AI. The McGovern Foundation is a nonprofit headquartered in Boston, United States, managing approximately USD 1.5 billion and focused on using AI and data technology to address public problems in climate, health, and education.

Dhar’s central argument rests on a distinction: legal obligation and trust-based responsibility are different things. “Of course there are responsibilities under the law, but there is also a responsibility of trust between companies and consumers,” he said. “Companies need to consider how the risk that an AI system poses to someone is monitored, and how to respond before harm occurs.” This means his standard goes beyond compliance—not merely satisfying what the law requires, but actively monitoring AI risks to consumers before damage materializes.

로그인 / 회원가입
독서리서치레이더투자 프레임워크

He offered a concrete example: AI analyzing health data and then offering different prices or service terms to different users. Preventing data leaks is only the baseline; what matters more is whether the AI’s judgments based on that data disadvantage the consumer. “The emergence of AI does not change people’s fundamental rights,” Dhar said. “When you make a transaction, you should be confident that you are treated equally as others, and that a company will not change your price because of information it holds about you.”

On the role of government, Dhar’s position is equally clear: companies cannot guarantee AI safety through self-regulation alone. “In South Korea as well, there should be public institutions that evaluate what companies are doing, and the government should have the power to halt operations when necessary,” he said. He argued that the pace of technological development far outstrips institutional response, and governments must accelerate. He also pointed out that governments worldwide have already articulated strong technical visions, and the next step is to develop those into a shared vision of what kind of society AI should shape, with participation from young people and other members of society.

Dhar himself did not provide a specific legal pathway for allocating responsibility, nor did he explain how “trust responsibility” could be translated into an enforceable regulatory standard; he called it “a very big question we must answer.”

Dhar visited South Korea to deliver a keynote at the Seoul Metropolitan Government’s Smart Life Week market forum, held alongside AI Festa at COEX in Seoul. The interview was prompted by ZDNet Korea’s earlier reporting on customer information misuse by the food-delivery platform Baedal Minjok. Dhar noted that the privacy and responsibility issues raised in that coverage were precisely the areas companies must examine first when introducing AI.

출처

  • zdnet.co.kr↗
주제

AI责任归属之争

当AI系统造成伤害时,谁该负责?美国麦戈文基金会总裁Billas Dhar指出,企业即便外包或使用第三方AI,对消费者的责任不可推卸,但开发方、部署方与第三方之间的责任划分仍是未解之谜。【背景】这一困境源于AI价值链的复杂性:数据提供商、基…

3 단계1 건업데이트됨 2026-10-20 18:50

현재까지의 과정

  1. 2025-04-04报告《AI价值链中的责任》发布,分析欧盟《人工智能责任指令》(已撤回)与修订后的《产品责任指令》,提出以过错责任为基线、高风险场景适用严格责任的框架。aigl.blog ↗
  2. 2026-01-19论文《无界伤害,有界法律》提交,指出领土受限的责任制度无法应对跨境AI风险,建议参考疫苗伤害和核责任体系,建立全球AI问责与赔偿架构。arxiv.org ↗
  3. 2026-10-20美国麦戈文基金会总裁Billas Dhar在首尔接受采访,强调企业使用AI时的消费者责任超越法律合规,需包含事前风险监控,并呼吁政府具备评估和暂停不安全AI使用的能力。zdnet.co.kr ↗

아직 해결되지 않은 문제

  • 在缺乏统一国际法的情况下,如何有效执行针对跨境AI伤害的赔偿责任?
  • 现有的科技错误与遗漏保险(Tech E&O)是否真正覆盖算法偏见、幻觉等特定AI风险,还是存在大量除外条款?
  • 政府应如何在促进创新与实施强制性安全评估之间取得平衡,特别是在物理AI即将大规模普及的背景下?
전체 주제 읽기 →
리서치 계속하기 →