2024年5月8日，新加坡人力部长陈诗龙医生在国会书面答复裕廊集选区议员陈有明，关于如何确保雇主使用的人工智能系统不会导致歧视性招聘行为的问题。

以下内容为新加坡眼根据国会英文资料翻译整理：

         确保雇主使用的人工智能系统不会导致歧视性招聘行为 

陈有明（裕廊集选区议员）询问人力部长：

(a) 有哪些保障措施可以确保雇主使用的人工智能 (AI) 系统不会导致不公平或歧视性的招聘行为？

(b) 可以采取哪些措施来确保受人工智能影响的招聘行为和人力资源决策的责任，包括将人力资源分析外包给使用人工智能的承包商？

陈诗龙医生(人力部长）：劳资政公平雇佣指导原则（Tripartite Guidelines on Fair Employment Practices）和即将出台的职场公平立法（Workplace Fairness Legislation）保护雇员免受工作场所歧视。如果雇主故意使用人工智能（AI）做出歧视性决定，受影响的员工将受到保护，免受工作场所歧视，人力部（MOM）也可以对雇主采取行动。

与此同时，在招聘和其他人力资源决策中使用人工智能可能会无意中导致歧视性的结果。为应对这一风险，政府已采取措施建立一个可信的环境，让人们相信人工智能的使用本身不会增加工作场所歧视的风险。这包括促进开发人工智能监管模式框架（Model AI Governance Framework），以指导企业开发和部署人工智能，并提供实际支持，如开源测试工具包，以根据国际公认的治理原则（包括公平性）验证人工智能系统的性能。如果某些人工智能系统导致歧视性就业决定，可向劳资政公平与良好雇佣联盟（Tripartite Alliance for Fair and Progressive Employment Practices）报告，该联盟将与雇主合作，确保此类系统的使用符合工作场所公平原则。

以下是英文质询内容：

ENSURING ARTIFICIAL INTELLIGENCE SYSTEMS USED BY EMPLOYERS DO NOT RESULT IN DISCRIMINATORY HIRING PRACTICES

 Dr Tan Wu Meng asked the Minister for Manpower (a) what are the safeguards to ensure that artificial intelligence (AI) systems used by employers do not result in unfair or discriminatory hiring practices; and (b) what measures can be taken to ensure accountability for hiring practices and human resource decisions which are influenced by AI, including the outsourcing of human resource analytics to contractors that use AI.

Dr Tan See Leng: The Tripartite Guidelines on Fair Employment Practices and the upcoming Workplace Fairness Legislation protect employees against workplace discrimination. If an employer intentionally uses artificial intelligence (AI) to make a discriminatory decision, the affected employee will be protected from discrimination in the workplace and the Ministry of Manpower (MOM) can take action against the employer.

At the same time, the use of AI in hiring and other human resource decisions may unintentionally result in outcomes that are discriminatory. To address this risk, the Government has implemented measures to establish a trusted environment where people can have confidence that the use of AI itself does not add to the risk of discrimination in the workplace. This includes facilitating the development of Model AI Governance Frameworks to guide businesses in the development and deployment of AI and providing practical support, such as open-source testing toolkits, to validate the performance of AI systems against internationally recognised governance principles, including fairness. If certain AI systems result in discriminatory employment decisions, they can be reported to the Tripartite Alliance for Fair and Progressive Employment Practices, which will work with the employer to ensure that the use of such systems are in line with the principles of workplace fairness.

HQ丨编辑

HQ丨编审

新加坡国会丨来源