数据更新: 2026-06-09 181 条辩论记录
来源: SPRS Hansard · PAIR Search 中文摘要由 AI 辅助生成
🏛️

国会 AI 焦点

新加坡议会关于人工智能的第一手辩论记录,含 AI 提炼的中文摘要、议员立场分析与政策规律洞察。

先找到你要的辩论

先搜索,再按类型、议题、议员、年份或争议度收窄。

快捷筛选

按议题浏览

先从高价值入口进入完整档案。

档案概览

点击柱状图筛选该年。

记录总数
181
覆盖年份
2015-2026
更新日期
2026-06-09
类型分布
书面答复 63 / 动议 7 / 口头答复 52 / 预算辩论 59
历年辩论趋势

辩论记录

列表默认压缩;展开卡片或进入详情页看完整摘要。

第 15 届国会 信息发布

过去三年 AI 诱发精神病的记录病例及应对 AI 相关心理健康问题的措施

Recorded Incidences of AI-Induced Psychosis in Past Three Years and Measures to Address AI-Related Mental Health Issues

工人党议员何廷儒(He Ting Ru)书面质询社会政策统筹部长兼卫生部长:卫生部是否注意到"AI 诱发精神病"(AI-induced psychosis)记录病例上升;若有,过去三年各年病例数为多少;以及部门如何与相关伙伴合作,更好地研究和应对这一问题。卫生部长王乙康(Ong Ye Kung)答复:国际上确有关于"密集使用人工智能可能引发心理紊乱"的报道,但这尚未成为确立的临床诊断;本地除少数零星个案外,并未观察到此类病例上升的明确趋...

政策信号: 新加坡卫生部门尚未把"AI 诱发精神病"纳入正式统计或诊断框架,采取"监测优先、暂不立规"的姿态,把 AI 心理健康风险定位为需持续观察的新兴议题,而非既成的公共卫生问题。
展开摘要、立场与原文

工人党议员何廷儒(He Ting Ru)书面质询社会政策统筹部长兼卫生部长:卫生部是否注意到"AI 诱发精神病"(AI-induced psychosis)记录病例上升;若有,过去三年各年病例数为多少;以及部门如何与相关伙伴合作,更好地研究和应对这一问题。卫生部长王乙康(Ong Ye Kung)答复:国际上确有关于"密集使用人工智能可能引发心理紊乱"的报道,但这尚未成为确立的临床诊断;本地除少数零星个案外,并未观察到此类病例上升的明确趋势。不过他强调,AI 与数字媒体可能带来的健康风险是"真实的关切",卫生部正对此密切监测。这一问答把"AI 心理健康风险"首次摆上国会台面:反对党推动政府正视并量化新兴风险,政府则以"尚非临床诊断、本地无明显趋势"审慎回应,同时不否认风险、保留持续监测的空间。

要点
  • • "AI 诱发精神病"目前尚非确立的临床诊断
  • • 本地除少数零星个案外,未见此类病例上升的明确趋势
  • • 卫生部将 AI 与数字媒体的潜在健康风险视为"真实关切",正密切监测
政府立场
卫生部对"AI 诱发精神病"持审慎观察态度:不将其认定为已确立的临床诊断,强调本地无明确上升趋势,但承认 AI 与数字媒体的心理健康风险真实存在,将持续密切监测。
质询立场
工人党何廷儒推动政府正视并量化 AI 相关精神健康风险,要求披露过去三年的病例数据,并说明跨伙伴的研究与应对机制。
"国际上虽有关于密集使用 AI 可能引发心理紊乱的报道,但这并非确立的临床诊断……我们正密切监测,因为 AI 与数字媒体带来的潜在健康风险确实值得关切。"
原文节选(英文)
Ms He Ting Ru asked the Coordinating Minister for Social Policies and Minister for Health three related questions: whether the Ministry is aware of increasing recorded incidences of artificial intelligence (AI)-induced psychosis; if so, the number of cases in each of the past three years; and how the Ministry is working with relevant partners to better study and tackle this issue. Mr Ong Ye Kung replied that there are international reports about psychological disturbances potentially related to intensive AI use, but that this is not an established clinical diagnosis. Other than a few anecdotal cases, no clear trends that such cases are on the rise have been observed locally. Nonetheless, the Minister said the Ministry is monitoring the matter closely, because the potential health risk posed by AI and digital media is of real concern.
第 15 届国会 信息发布

one-north 人工智能园区的能耗预测与能效措施

Energy Consumption Projections of AI Park at one-north and Energy Efficiency Measures

工人党议员陈立峰(Dennis Tan)就 one-north 新建人工智能(AI)园区的资源消耗,书面质询数字发展与信息部:园区满负荷运营时的年度电力与水耗预测为多少;部门如何衡量国家 AI 计划(National AI Missions)下企业的软件级能效;以及政府是否会强制大型 AI 模型披露能耗强度,以缓解对全国能源需求的反弹效应。部长杨莉明(Josephine Teo)答复:one-north 的 AI 园区并非数据中心园区,...

政策信号: 新加坡倾向以行业国际标准、绿色软件指南与《数字基础设施法案》的数据中心能效门槛来约束 AI 能耗,监管重心落在基础设施层而非模型层,暂不对单个大型模型施加强制能耗披露义务。
展开摘要、立场与原文

工人党议员陈立峰(Dennis Tan)就 one-north 新建人工智能(AI)园区的资源消耗,书面质询数字发展与信息部:园区满负荷运营时的年度电力与水耗预测为多少;部门如何衡量国家 AI 计划(National AI Missions)下企业的软件级能效;以及政府是否会强制大型 AI 模型披露能耗强度,以缓解对全国能源需求的反弹效应。部长杨莉明(Josephine Teo)答复:one-north 的 AI 园区并非数据中心园区,而是面向不断壮大的 AI 开发者与解决方案商社群的聚集地,预计其年度电力与水耗将与榜鹅数码园区等技术密集型工业园相当。她强调新加坡致力于可持续地推进 AI 能力——国家 AI 研发计划已将"资源高效型 AI"列为研究重点;绿色 AI 软件目前仍属新兴领域、尚无国际公认标准,信息通信媒体发展局(IMDA)正与绿色软件基金会合作制定新的"软件碳强度"ISO 标准方法论,并通过"绿色软件试验"与新发布的《企业实用绿色软件指南》推动算力的高效利用;数据中心方面则依托 IMDA 的"绿色数据中心路线图",部门将通过正在与业界磋商制定的《数字基础设施法案》引入数据中心基线能效标准。

要点
  • • one-north AI 园区定位为 AI 开发者与方案商聚集地,非数据中心园区,能耗预计与榜鹅数码园区相当
  • • 国家 AI 研发计划已将"资源高效型 AI"列为研究重点
  • • IMDA 正与绿色软件基金会合作制定"软件碳强度"ISO 标准,并推《企业实用绿色软件指南》
  • • 政府将通过《数字基础设施法案》为数据中心引入基线能效标准
政府立场
政府认为 one-north AI 园区能耗可控(对标现有技术密集型园区),主张以"资源高效型 AI"研究、绿色软件国际标准与《数字基础设施法案》的数据中心能效门槛来管理 AI 的能源足迹,暂不单独强制大型模型披露能耗强度。
质询立场
工人党陈立峰推动量化披露:要求公开园区满负荷能耗预测、建立软件级能效衡量方法,并强制大型 AI 模型披露能耗强度,以防 AI 扩张抬高全国能源需求的反弹效应。
"one-north 的人工智能园区并非数据中心园区,而是面向不断壮大的 AI 开发者与解决方案商社群的聚集地。"
原文节选(英文)
Mr Dennis Tan Lip Fong asked the Minister for Digital Development and Information three related questions about the new artificial intelligence (AI) park at one-north: the projected annual electricity and water consumption at full capacity; how the Ministry would measure software-level energy efficiency for firms under the national AI Missions; and whether the Government would mandate energy-intensity disclosures for large-scale AI models to mitigate any rebound effect on national energy demand. Mrs Josephine Teo replied that the planned AI park is not a data centre park but is designed to be a hub for a growing community of AI developers and solution providers, so its annual electricity and water consumption is expected to be comparable to other technology-intensive industrial estates such as the Punggol Digital District. She added that Singapore is committed to advancing its AI capabilities sustainably: under the National AI Research and Development plan, resource-efficient AI has been identified as a research priority, which together with close monitoring of practices elsewhere will inform appropriate policy measures. On software, she noted that green or energy-efficient AI software is still a nascent field with no internationally recognised standard today; the Infocomm Media Development Authority (IMDA) is working closely with the Green Software Foundation on the methodology for the new Software Carbon Intensity ISO standard, and in the meantime promotes the efficient use of computing resources through initiatives such as the Green Software Trials and the recently launched Practical Green Software Guide for Enterprises. On data centres, she said steps had been taken to improve sustainability through IMDA's Green DC roadmap, and that the Ministry will introduce baseline energy-efficiency standards for data centres through the Digital Infrastructure Act, which is being developed in consultation with the industry.
第 15 届国会 信息发布

自动驾驶车辆商业化上路前的可量化安全指标及规范AI决策的标准

Quantifiable Safety Metrics and AV Standards Governing AI Decision-Making Prior to Commercial Deployment on Public Roads

议员Alex Yeo书面询问代交通部长:陆路交通管理局(LTA)的两阶段部署就绪评估中是否有自动驾驶车辆(AV)必须达到的具体、可量化的安全指标;这些标准是什么;以及在首批商业AV获准上路前,《技术参考68》(TR 68)是否会进一步加强,或是否会出台一套规范AI决策协议的正式AV标准。代部长Jeffrey Siow答复:AV部署就绪评估框架规定了技术性能和公众接受度指标,AV须先达标才能载客并逐步过渡到无人驾驶运营,具体包括在实际部署...

政策信号: 新加坡的AV监管走渐进路线:以个案授权和性能指标把关,而非预先制定AI决策的硬性标准,TR 68复审预示商业化前规则将更新但节奏由监管方掌控。
🎙️ Mr Alex Yeo · Mr Jeffrey Siow
展开摘要、立场与原文

议员Alex Yeo书面询问代交通部长:陆路交通管理局(LTA)的两阶段部署就绪评估中是否有自动驾驶车辆(AV)必须达到的具体、可量化的安全指标;这些标准是什么;以及在首批商业AV获准上路前,《技术参考68》(TR 68)是否会进一步加强,或是否会出台一套规范AI决策协议的正式AV标准。代部长Jeffrey Siow答复:AV部署就绪评估框架规定了技术性能和公众接受度指标,AV须先达标才能载客并逐步过渡到无人驾驶运营,具体包括在实际部署路线上累计足够的无需安全员干预的行驶里程,以及在授权地理围栏区域内应对各种交通场景的能力。作为AV行业指南的《技术参考68》上次更新是在2021年,目前正在复审;在此期间,AV部署将继续沿用现行授权制度。答复未列出具体量化阈值,也未承诺在商业部署前出台规范AI决策的正式标准。

要点
  • • 部署就绪评估含技术性能与公众接受度两类指标
  • • 要求在部署路线累计足够的无安全员干预里程
  • • 《技术参考68》2021年最后更新,目前正在复审
  • • 复审期间AV部署沿用现行授权制度
政府立场
政府认为现行部署就绪评估框架和授权制度足以保障AV安全,《技术参考68》正在复审,暂不另行出台规范AI决策的正式标准。
"作为新加坡AV行业指南的《技术参考68》上次更新是在2021年,目前正在复审。"
原文节选(英文)
Mr Alex Yeo asked the Acting Minister for Transport (a) whether the Land Transport Authority (LTA) has specific, quantifiable safety metrics in the two-stage Deployment Readiness Assessment that autonomous vehicles (AVs) must meet; (b) if so, what these standards are; and (c) whether Technical Reference 68 will be further enhanced, or a formal set of AV standards that governs artificial intelligence decision-making protocols will be introduced, before the first commercial AVs are allowed on the roads. Mr Jeffrey Siow replied that the deployment readiness assessment framework for autonomous vehicles sets out technical performance and public acceptance metrics that must be met before the AVs are allowed to take passengers and eventually progress to driverless operations. These include clocking sufficient distance within the actual deployment route without the need for intervention by safety operators, and the ability to handle various traffic scenarios presented within the authorised geo-fenced area. On standards, the Acting Minister said that Technical Reference 68, which serves as guidelines for the AV industry in Singapore, was last updated in 2021 and is currently being reviewed. In the meantime, AV deployments will continue under the existing authorisation regime. The written answer did not enumerate specific quantitative thresholds for the technical performance or public acceptance metrics, and did not commit to introducing a formal set of AV standards governing AI decision-making protocols before the first commercial AVs are allowed on public roads.
第 15 届国会 信息发布

学校使用AI工具的数据保护标准及AI对学生学习成效影响的评估

Data Protection Standards for AI Tools Usage in Schools and Assessing AI's Impact on Student Learning Outcomes

议员Eileen Chong Pei Shan就学校使用AI工具向教育部长提出两项书面询问:一问学校在指导学生使用非新加坡学生学习平台(SLS)托管的AI工具时须遵循哪些最低数据保护标准、学生为完成课业使用此类工具前是否需要家长知情或同意、以及如何监督合规情况;二问教育部关于AI对学生学习影响研究的范围(覆盖的教育阶段和测量的学习成效指标)、研究结果的发布时间表,以及结果将如何用于制定或修订AI教育战略。教育部长李智陞的书面答复未展开实...

政策信号: 学校AI工具的数据保护与学习成效评估已成为国会高频议题,教育部选择以统一口径集中回应,显示其AI教育政策仍在研究先行、谨慎披露的阶段。
🎙️ Ms Eileen Chong Pei Shan · Desmond Lee
展开摘要、立场与原文

议员Eileen Chong Pei Shan就学校使用AI工具向教育部长提出两项书面询问:一问学校在指导学生使用非新加坡学生学习平台(SLS)托管的AI工具时须遵循哪些最低数据保护标准、学生为完成课业使用此类工具前是否需要家长知情或同意、以及如何监督合规情况;二问教育部关于AI对学生学习影响研究的范围(覆盖的教育阶段和测量的学习成效指标)、研究结果的发布时间表,以及结果将如何用于制定或修订AI教育战略。教育部长李智陞的书面答复未展开实质内容,仅指这些问题已由教育部在2026年5月6日对第2至5号口头议会询问的答复中一并处理,并请议员参阅《官方报告》第96卷第30期口头答复部分"监测小学生使用AI"一节。该答复反映家长关切的数据保护与学习成效问题已在前一日的口头质询中集中回应。

要点
  • • 议员追问非SLS托管AI工具的最低数据保护标准与家长同意要求
  • • 同时询问AI学习影响研究的范围、发布时间表和战略用途
  • • 教育部长书面答复仅援引5月6日口头答复,未给新内容
  • • 相关回应见《官方报告》第96卷第30期"监测小学生使用AI"
政府立场
教育部认为相关数据保护与AI学习影响问题已在2026年5月6日对口头询问的集中答复中处理,书面答复仅作援引而不再另行展开。
"这一问题已由教育部对2026年5月6日第2至5号口头议会询问的答复处理。"
原文节选(英文)
Ms Eileen Chong Pei Shan asked the Minister for Education two written questions on the use of AI tools in schools. First, she asked (a) what minimum data protection standards schools are required to apply when directing students to use AI tools not hosted on the Singapore Student Learning Space; (b) whether parental notification or consent is required before students use such tools for school-assigned work; and (c) how compliance with these standards is monitored. Second, she asked (a) what is the scope of the Ministry's study on the impact of artificial intelligence (AI) on students' learning, in terms of education levels covered and learning outcomes measured; (b) what is the expected timeline for findings to be published; and (c) how the findings will be used to inform or revise the AI in education strategy. Mr Desmond Lee replied that these questions had been addressed by the Ministry of Education's answer to oral Parliamentary Question Nos 2 to 5 on 6 May 2026, and referred the Member to "Monitoring AI Use by Primary School Students", Official Report, 6 May 2026, Vol 96, Issue 30, Oral Answers to Questions section. The written reply itself provided no further substantive details on data protection standards, parental consent requirements, compliance monitoring, or the scope and timeline of the Ministry's study; the full government response on these matters is recorded under the oral answer delivered the previous day.
第 15 届国会 实质辩论

人工智能(AI)转型不容无就业式增长(主辩论)

An Artificial Intelligence (AI) Transition with No Jobless Growth (Main Debate)

5月6日国会续辩职总秘书长黄志明提出的"人工智能转型不容无就业式增长"动议,约20名议员发言,是第15届国会迄今最具分量的AI辩论。动议要求承认AI的经济变革力量,把增长锚定在公平、韧性与人人有机会,装备工人与企业,确认新加坡不容无就业增长。执政党与工运议员聚焦岗位重设计、公司培训委员会(CTC)与新设的三方就业理事会;工人党议员全数支持动议但另提结构方案:严燕松倡全民AI股权基金(成年公民每年500新元分红加在职培训基金),Andre...

政策信号: AI就业冲击已升为国家级议程:政府明确以三方机制而非立法保障应对AI转型,公共资助将与工人成果挂钩,求职者援助计划门槛与裁员通报制度料将检讨增强。
展开摘要、立场与原文

5月6日国会续辩职总秘书长黄志明提出的"人工智能转型不容无就业式增长"动议,约20名议员发言,是第15届国会迄今最具分量的AI辩论。动议要求承认AI的经济变革力量,把增长锚定在公平、韧性与人人有机会,装备工人与企业,确认新加坡不容无就业增长。执政党与工运议员聚焦岗位重设计、公司培训委员会(CTC)与新设的三方就业理事会;工人党议员全数支持动议但另提结构方案:严燕松倡全民AI股权基金(成年公民每年500新元分红加在职培训基金),Andre Low主张无收入上限的裁员保险、再培训税收抵免与"AI收益审计",Kenneth Tiong呼吁高端AI工具全民可及。人力部长陈诗龙驳斥工人党方案是"和解金"而非赋权,引人力部调查称采用AI企业仅约6%裁员,承诺研究提高求职者援助收入门槛并推动更早裁员通报。朝野最终均表态支持动议。

要点
  • • 约20名议员发言近7小时,朝野全数表态支持动议,但在政策工具上针锋相对
  • • 工人党提全民AI股权基金(成人每年500新元分红)与无收入上限裁员保险
  • • 陈诗龙引人力部调查:采用AI企业仅约6%裁员,七成已见生产力提升
  • • 政府允研究提高求职者援助收入门槛,推动雇主在员工离职前通报裁员
政府立场
政府拒绝以现金再分配为主的保障路线,坚持"投资于人":通过技能与劳动力发展局、三方就业理事会和企业劳动力转型配套让工人随AI增长共同进步,承诺"保不住每一份工作,但会保护每一个工人"。
质询立场
工人党议员支持动议本身,但批评现行求职者援助计划门槛过低、递减设计逼工人仓促就业,主张无收入上限的裁员保险、全民AI股权基金、再培训税收抵免与强制性AI转型通知等结构性保障。
"你们两位的提案都不是赋权。在我看来,那是一种和解金——认定大规模岗位流失不可避免,我们能做的最多只是减轻冲击。"
原文节选(英文)
Parliament resumed debate on NTUC Secretary-General Ng Chee Meng's motion calling on the House to recognise AI's transformative power, anchor AI-enabled growth in fairness, resilience and opportunity for all, equip workers and enterprises, and affirm that Singapore must not have jobless growth because every worker matters. Around twenty MPs spoke. NMP Mark Lee urged a single front door for enterprise AI grants and a shift to place-and-train; Mr Saktiandi Supaat warned of a divide between AI haves and have-nots. Ms Yeo Wan Ling said union ground work showed bus captains spend only about 20% of their day driving, so job redesign must rest on actual workflows. Workers' Party MPs supported the motion while proposing alternatives. Mr Gerald Giam proposed a National AI Equity Fund paying a $500 annual dividend to every adult citizen plus an on-the-job mastery fund, financed by a corporate tax rise on firms with profits above $100 million and a larger draw on net investment returns. Mr Andre Low argued the tapering, $5,000-capped Jobseeker Support scheme pushes workers into the first job rather than the right one, urging WP's redundancy insurance paying 40% of last-drawn salary with no ceiling, a retraining tax credit and an annual AI gains audit. Mr Kenneth Tiong called for universal premium AI agent access and a 90-day transition notice. Ms He Ting Ru cited IMF estimates that 77% of Singapore's workforce is highly exposed to AI; Assoc Prof Jamus Lim warned of a graduate hiring slowdown. Labour MP Patrick Tay urged raising the Jobseeker Support ceiling to the PME median of $8,400. Responding, Minister of State Jasmin Lau said companies receiving public support must deliver worker outcomes; Senior Minister of State Desmond Tan detailed NTUC's AI-Ready SG, with 3,800 Company Training Committees and 50% subsidies on 21 AI tools. Manpower Minister Tan See Leng rejected the WP's payout-centred proposals as a settlement rather than empowerment, citing MOM's survey that only about 6% of AI-adopting firms cut headcount; he committed to study raising the Jobseeker Support threshold, sought retrenchment notification by workers' last day, and confirmed six months of free premium AI tools with selected SkillsFuture courses. Every speaker, including all WP members, declared support for the motion.
第 15 届国会 实质辩论

不容无就业增长的人工智能(AI)转型(辩论总结)

An Artificial Intelligence (AI) Transition with No Jobless Growth (Debate Conclusion)

"AI转型不容无就业增长"动议辩论5月6日进入收尾。澄清环节中,工人党议员刘吴扬强调强社会安全网与鼓励国人拥抱AI并非零和;严燕松为其"全国AI公平基金"辩护,称该基金不是"对失败的补偿"、近半资金将直接投资工人技能,并追问政府如何让工人结构性分享AI生产力红利,人力部长陈诗龙回应分享方式是"实际收入的提升",新加坡劳动力发展局(SWDA)将设明确KPI,把企业劳动力培训配套与职位重设计同实际工资和职业发展挂钩。工人党Kenneth T...

政策信号: 动议一致通过确立"AI增长不容无就业化"为跨党派共识,但政府偏好三方协作与培训路径、回绝结构性财富分享机制的立场,与反对党的制度化保障主张之间的张力,将持续塑造就业法修订与AI补助计划的设计。
展开摘要、立场与原文

"AI转型不容无就业增长"动议辩论5月6日进入收尾。澄清环节中,工人党议员刘吴扬强调强社会安全网与鼓励国人拥抱AI并非零和;严燕松为其"全国AI公平基金"辩护,称该基金不是"对失败的补偿"、近半资金将直接投资工人技能,并追问政府如何让工人结构性分享AI生产力红利,人力部长陈诗龙回应分享方式是"实际收入的提升",新加坡劳动力发展局(SWDA)将设明确KPI,把企业劳动力培训配套与职位重设计同实际工资和职业发展挂钩。工人党Kenneth Tiong质疑技能创前程课程质量、主张普及前沿AI工具;Mark Lee则质疑其"AI裁员90天强制通知"提案在渐进式转型下难以操作。杨涴凌追问AI补助是否强制绑定职位重设计,政务部长Jasmin Lau表示方向已定、细节将与三方伙伴商定。黄志明总结这场历时7小时18分钟、24场发言的辩论,支持把求职者援助计划合格门槛升至PME中位月入约8,400新元,表明职总对援助形式持开放态度,动议最终获国会一致通过。

要点
  • • 辩论历时7小时18分钟、24场发言,动议最终获国会一致通过
  • • 陈诗龙:SWDA将设KPI,把培训配套与职位重设计同实际工资挂钩
  • • 工人党坚持AI公平基金、裁员保险、90天AI裁员通知等结构性方案
  • • 职总支持把JSS合格门槛升至PME中位月入约8,400新元
政府立场
政府承诺不把AI增长的结果交给市场决定:公共支持将附带工人结果预期,并通过SWDA的KPI让AI红利以实际收入提升和职业发展的形式分享给工人,而非建立结构性财富再分配基金。
质询立场
工人党议员主张更结构性的保障——全国AI公平基金、裁员保险、AI裁员90天强制通知与可申诉机制——并认为强社会安全网正是国人敢于冒险拥抱AI机遇的前提,而非与之对立。
"全国AI公平基金不是对失败的补偿,而是为安心与成功提供所需的保障。"
原文节选(英文)
Debate on the Motion resumed on 6 May 2026 with a round of clarifications. Mr Andre Low (WP) welcomed Minister of State Jasmin Lau's position that automation and augmentation are not mutually exclusive, and pressed Minister Tan See Leng on his characterisation of WP proposals, arguing that strong social safety nets are precisely what enable Singaporeans to take risks and embrace AI. Mr Gerald Giam (WP) defended his proposed National AI Equity Fund — one limb giving workers a direct stake in prosperity, the other, nearly half the fund, investing in skills and on-the-job training — and asked how the Government would move from discretionary spending to a structural sharing of AI wealth. Dr Tan See Leng replied that sharing would come through real income improvement, with SWDA setting clear KPIs tying the Enterprise Workforce Training Package and job redesign to real wages and career progression. Mr Kenneth Tiong (WP) questioned whether SkillsFuture enrolment is a reliable proxy for seriousness, citing residents unimpressed with course quality; Dr Tan replied that the aim is to bring as broad a segment of the population as possible to "base camp". Ms Yeo Wan Ling asked whether job redesign would be an explicit condition of AI grants and what worker outcomes would be tracked; Ms Lau said the direction is committed but details would be discussed with tripartite partners. Mr Mark Lee challenged Mr Tiong's proposed 90-day mandatory notice for AI-driven role elimination as operationally ambiguous when transformation is gradual and task-based; Mr Tiong pointed to an adjudicative process, citing a court case in Hangzhou. Wrapping up after 24 speeches and over seven and a half hours, Mr Ng Chee Meng addressed Members' contributions across his four practical moves, supported raising Jobseeker Support eligibility to the PME median income of about $8,400 as called for by Mr Patrick Tay, said NTUC is not wedded to any particular form of support in response to Mr Low's redundancy insurance proposal, and declared that "this House stands with you". After a final exchange with Mr Giam on wage support for on-the-job training, the question was put and agreed to.
第 15 届国会 信息发布

AI 研发药物的监管框架及其对临床试验的影响,以及国家患者数据保护措施是否充分

Regulatory Framework for AI-developed Drugs and Implications on Clinical Trials, and Adequacy of Data Protection Safegua...

叶汉荣议员书面询问社会政策统筹部长兼卫生部长:卫生部是否研究过 AI 研发药物能否缩短或绕过临床试验;若可以,AI 药物的审批与传统药物有何不同;现行哪些法规管辖 AI 医疗创新;以及当 AI 为产品概念化而调用国家患者数据时,现有个人数据保护与网络安全措施能否防止数据泄露。卫生部长王乙康答复:AI 研发药物与传统药物必须满足同样的国际质量、安全和疗效标准,不存在捷径;卫生科学局的监管思路与美国 FDA、欧洲药品管理局等国际机构一致,这...

政策信号: 新加坡在 AI 医药监管上选择与 FDA、EMA 等国际机构对齐的"同一标准"路线,强调 AI 提升研发效率但不松动安全证据门槛,体现稳健而非激进的医疗 AI 监管取向。
展开摘要、立场与原文

叶汉荣议员书面询问社会政策统筹部长兼卫生部长:卫生部是否研究过 AI 研发药物能否缩短或绕过临床试验;若可以,AI 药物的审批与传统药物有何不同;现行哪些法规管辖 AI 医疗创新;以及当 AI 为产品概念化而调用国家患者数据时,现有个人数据保护与网络安全措施能否防止数据泄露。卫生部长王乙康答复:AI 研发药物与传统药物必须满足同样的国际质量、安全和疗效标准,不存在捷径;卫生科学局的监管思路与美国 FDA、欧洲药品管理局等国际机构一致,这些机构已就 AI 在药物研发中的负责任使用提出关键原则。患者数据受到包括《个人数据保护法》在内的现有数据保护和网络安全措施的有力保护,政府将持续监测发展并按需强化保障。答复确认了"同一标准"原则——AI 不改变药物审批的证据门槛。

要点
  • • 叶汉荣问 AI 研发药物能否缩短或绕过临床试验
  • • 王乙康:AI 药物与传统药物须满足同一国际质量安全疗效标准
  • • 卫生科学局监管思路与美国 FDA、欧洲药品管理局对齐
  • • 患者数据受《个人数据保护法》等现有保障有力保护
政府立场
AI 研发药物不享有审批捷径,必须与传统药物满足同一国际质量、安全与疗效标准,患者数据已有有力保护并将按需强化。
"无论是 AI 研发还是传统研发的药物,都必须满足同样的国际质量、安全和疗效标准。"
原文节选(英文)
Mr Yip Hon Weng asked the Coordinating Minister for Social Policies and Minister for Health whether the Ministry has studied if AI-developed drugs can shorten or bypass clinical trials; if so, how the regulatory approval of AI-developed drugs would differ from conventional products; what regulations currently govern AI healthcare innovations; and whether existing personal data protection and cybersecurity safeguards can prevent data leakage when AI accesses national patient data for product conceptualisation. Mr Ong Ye Kung, the Minister for Health, replied that both artificial intelligence-developed and conventionally developed drugs must meet the same international standards of quality, safety and efficacy. The Health Sciences Authority's regulatory approach is aligned with that of international agencies, such as the US Food and Drug Administration and the European Medicines Agency, which have outlined key principles on the responsible use of AI in drug development. On the data protection questions, the Minister stated that patient data is robustly protected, including when it is used for AI development. Existing data protection and cybersecurity safeguards, including those under the Personal Data Protection Act, ensure that patient confidentiality is maintained and that the data is protected. He added that the Government will continue to monitor developments and strengthen its safeguards as needed. The reply gave no indication of any separate or expedited approval pathway for AI-developed drugs: the same evidentiary requirements, including clinical trials, apply regardless of how a drug candidate was discovered or designed, and Singapore's regulator will track how international counterparts evolve their principles for AI in drug development.
第 15 届国会 温和质询

深度伪造检测器 INDEPTH 面向公众开放前的准确率基准与责任框架,及在 ScamShield 中集成实时媒体验证

Accuracy Benchmarks and Liability Frameworks for Intelligence Deepfake Detector Before Public Rollout and Integrating Re...

工人党议员林瑞莲书面询问数字发展与信息部长:智能深度伪造检测器 INDEPTH 从公共部门使用转向公众使用前,须满足什么技术准确率基准或责任框架;政府是否会在 ScamShield 应用中试点验证 API,为公民提供可疑媒体的实时风险评分。部长杨莉明答复:INDEPTH 是专为政府机构设计的深度伪造检测平台,不打算向公众开放;公开其检测能力不符合公共利益,因为恶意行为者可能利用这些信息。关于 ScamShield,她说明该应用可拦截已核...

政策信号: 政府在深度伪造防御上坚持"能力保密"路线:检测工具留在政府内部、不向公众下放,公民端防护走 ScamShield 渐进强化与公众教育,而非提供检测即服务。
展开摘要、立场与原文

工人党议员林瑞莲书面询问数字发展与信息部长:智能深度伪造检测器 INDEPTH 从公共部门使用转向公众使用前,须满足什么技术准确率基准或责任框架;政府是否会在 ScamShield 应用中试点验证 API,为公民提供可疑媒体的实时风险评分。部长杨莉明答复:INDEPTH 是专为政府机构设计的深度伪造检测平台,不打算向公众开放;公开其检测能力不符合公共利益,因为恶意行为者可能利用这些信息。关于 ScamShield,她说明该应用可拦截已核实的诈骗电话、过滤诈骗短信,并允许用户查验和举报含深度伪造等篡改痕迹的可疑来电、消息及链接;政府将继续强化其检测能力和公众教育,但现阶段无计划开发实时风险评分的验证 API。张力在于:反对党要求把政府级检测工具下放给公众,政府以安全保密为由拒绝。

要点
  • • 林瑞莲问 INDEPTH 开放公众使用前须满足什么准确率与责任框架
  • • 杨莉明:INDEPTH 专供政府机构,公开检测能力恐被恶意利用
  • • ScamShield 已可举报含深度伪造痕迹的可疑消息和链接
  • • 现阶段无计划在 ScamShield 开发实时风险评分验证 API
政府立场
INDEPTH 定位为政府机构专用平台、不向公众开放,以免恶意者利用其检测能力信息;公众反诈依托 ScamShield 现有功能持续强化,暂不开发实时风险评分 API。
质询立场
工人党林瑞莲主张让公众也能获得深度伪造检测能力,要求明确 INDEPTH 公众化的技术与责任门槛,并建议在 ScamShield 试点实时验证 API。
"公开其检测能力不符合公共利益,因为恶意行为者可能利用这些信息。"
原文节选(英文)
Ms Sylvia Lim asked the Minister for Digital Development and Information what technical accuracy benchmarks or liability frameworks the Intelligent Deepfake Detector (INDEPTH) must meet before it is transitioned from public service use to public use, and whether the Government will pilot a verification API within the ScamShield app to provide citizens with real-time risk scores for suspicious media. Mrs Josephine Teo, the Minister for Digital Development and Information, replied that INDEPTH is a deepfake detection platform designed specifically for Government agencies and is not intended for public use. She stated that it is not in the public interest to reveal its detection capabilities, as malicious actors may exploit such information. On the second question, the Minister explained that the ScamShield app was developed by Open Government Products in collaboration with the National Crime Prevention Council and the Singapore Police Force to protect the public from scams. It blocks scam calls from numbers verified by the authorities and filters scam SMSes. It also allows users to check and report suspicious calls, messages — such as SMS, WhatsApp and Telegram — and website links, including those that may show signs of digital manipulation, such as deepfakes. The Government will continue to strengthen ScamShield's existing scam detection capabilities and reporting channels, and will step up public education to better protect users from evolving scam tactics. However, there are no plans at this juncture to develop a verification API within the ScamShield app to provide real-time risk scores for suspicious media.
第 15 届国会 温和质询

确保面向公众的自主 AI 代理具备实质性人类问责,及高风险行业转向强制治理的路径

Ensuring Meaningful Human Accountability for Public-facing Autonomous AI Agents and Pathways to Mandatory Governance in...

工人党议员林瑞莲就《代理式 AI 示范治理框架》发布后续书面询问数字发展与信息部长:在没有明确披露要求的情况下,部门如何确保与公众交互的自主 AI 代理具备"实质性人类问责";什么样的触发条件会促使该框架从自愿性准则转为高风险行业的强制性标准。部长杨莉明答复:框架为机构部署代理式 AI 提供指引——高风险或不可逆的操作不得在无人审核下执行,机构应设置需人工批准的检查点或行动边界;框架同时强调对用户透明,例如预先声明用户在与代理交互及其能...

政策信号: 新加坡对代理式 AI 延续"软法先行"路线:以示范框架设定人工审核与透明度预期,依靠行业监管机构和国际实践迭代,强制监管的门槛刻意留白,保留政策弹性。
展开摘要、立场与原文

工人党议员林瑞莲就《代理式 AI 示范治理框架》发布后续书面询问数字发展与信息部长:在没有明确披露要求的情况下,部门如何确保与公众交互的自主 AI 代理具备"实质性人类问责";什么样的触发条件会促使该框架从自愿性准则转为高风险行业的强制性标准。部长杨莉明答复:框架为机构部署代理式 AI 提供指引——高风险或不可逆的操作不得在无人审核下执行,机构应设置需人工批准的检查点或行动边界;框架同时强调对用户透明,例如预先声明用户在与代理交互及其能力和数据访问权限。她表示代理式 AI 的应用场景和保障仍在演变,政府将与行业监管机构观察各行业部署实践、借鉴国际经验、按需调整框架——未给出从自愿转向强制的触发条件。张力在于:反对党追问治理硬化的明确路径,政府保持"观察—调整"的弹性立场。

要点
  • • 林瑞莲追问自主 AI 代理如何落实"实质性人类问责"
  • • 框架要求高风险或不可逆操作必须经人工审核批准
  • • 框架强调预先告知用户在与 AI 代理交互及其权限
  • • 政府未给出自愿框架转为强制标准的具体触发条件
政府立场
代理式 AI 治理现阶段以自愿性框架引导——高风险操作须人工审核、对用户保持透明,政府与行业监管机构持续观察部署实践并按需调整,暂不设定转为强制标准的触发条件。
质询立场
工人党林瑞莲质疑无披露要求下"实质性人类问责"难以落实,要求政府明确框架从自愿准则升级为高风险行业强制标准的触发条件。
"不应允许高风险或不可逆的操作在没有人工审核的情况下执行。"
原文节选(英文)
Ms Sylvia Lim asked the Minister for Digital Development and Information, following the launch of the Model AI Governance Framework for Agentic AI, how the Ministry intends to ensure "meaningful human accountability" for autonomous AI agents that interact with the public in the absence of explicit disclosure requirements, and what triggers would necessitate transitioning the Framework from a voluntary code to enforceable standards for high-risk sectors. Mrs Josephine Teo, the Minister for Digital Development and Information, replied that the Model Artificial Intelligence Governance Framework for Agentic AI sets out guidance for organisations to ensure meaningful human accountability in deploying agentic AI. Organisations should not allow high-stakes or irreversible actions to take place without human review; appropriate actions therefore include identifying checkpoints or action boundaries that require human approval. The Framework also emphasises transparency towards users, such as declaring upfront that users are interacting with agents, as well as the agents' capabilities and data access. The Minister noted that agentic AI use cases and the appropriate safeguards are still evolving. Hence, together with sector regulators, the Government will continue to monitor how various sectors deploy agentic AI and put these principles into practice, and will continue to consult and learn from best practices internationally, making adjustments to the framework as necessary. The reply did not specify any concrete triggers that would convert the voluntary framework into enforceable standards for high-risk sectors, leaving the pathway from guidance to regulation open-ended.
第 15 届国会 温和质询

对AI生成的政治深度伪造视频的评估及POFMA与OCHA下的监管介入门槛

Assessment of AI-generated Deepfake Political Videos and Regulatory Intervention Thresholds under POFMA and OCHA

工人党议员林瑞莲书面询问数字发展与信息部长:政府在判定涉及政治公职人员的AI深度伪造视频"明显属于伪造"、无需援引《防止网络虚假信息和网络操纵法》(POFMA)发出指示之前,是否对弱势群体做过测试;以及此类视频须造成何种程度的公众混淆,才会触发《网络犯罪危害法》(OCHA)下的主动干预。部长杨莉明答复:POFMA下部长可对在新加坡传播且涉及公共利益的虚假事实陈述发出更正指示,包括深度伪造视频;OCHA下政府可在合理怀疑网络活动涉及附表一...

政策信号: 政府在深度伪造监管上坚持个案裁量而非量化门槛,并把媒体素养教育置于与立法同等的位置,短期内不会为干预设定硬性标准。
展开摘要、立场与原文

工人党议员林瑞莲书面询问数字发展与信息部长:政府在判定涉及政治公职人员的AI深度伪造视频"明显属于伪造"、无需援引《防止网络虚假信息和网络操纵法》(POFMA)发出指示之前,是否对弱势群体做过测试;以及此类视频须造成何种程度的公众混淆,才会触发《网络犯罪危害法》(OCHA)下的主动干预。部长杨莉明答复:POFMA下部长可对在新加坡传播且涉及公共利益的虚假事实陈述发出更正指示,包括深度伪造视频;OCHA下政府可在合理怀疑网络活动涉及附表一所列罪行、或为诈骗及恶意网络活动做准备时,指示平台屏蔽新加坡用户访问。她强调政府对有害内容做整体评估,干预不取决于单一因素;并称公共教育与立法同等重要,列举IMDA"数字技能为生活"框架和SG数字办公室面向年长者的生成式AI工作坊。

要点
  • • 林瑞莲问:判定深伪视频"明显伪造"前是否测试过弱势群体
  • • 杨莉明:POFMA可发更正指示,OCHA可令平台屏蔽访问
  • • 政府整体评估有害内容,干预不取决于任何单一因素
  • • IMDA数字技能框架与年长者生成式AI工作坊承担公共教育
政府立场
政府认为现有POFMA与OCHA框架足以应对深度伪造视频,干预基于个案整体评估而非单一量化门槛,公共教育与立法同等重要。
质询立场
林瑞莲质疑政府在不动用POFMA时是否实际验证过弱势群体的辨别能力,并要求政府明确OCHA主动干预所需的公众混淆门槛。
"因此,公共教育举措与立法同等重要,有助于我们的人民在上网时更加警惕、更有辨别力。"
原文节选(英文)
Ms Sylvia Lim asked the Minister for Digital Development and Information (a) whether the Government does testing on vulnerable groups before determining that AI-generated deepfake videos involving political office-holders are so obviously fabricated that directions under the Protection from Online Falsehoods and Manipulation Act 2019 (POFMA) are unnecessary; and (b) what threshold of public confusion arising from such videos must be met before a proactive intervention under the Online Criminal Harms Act 2023 (OCHA). Mrs Josephine Teo replied that under POFMA, a Minister can issue a Correction Direction if a false statement of fact has been communicated in Singapore and it is in the public interest to issue it, and that this includes online falsehoods communicated via deepfake videos. Under OCHA, the Government can issue directions to online service providers to disable Singapore users' access to online content or activity when there is reasonable suspicion that the activity is in furtherance of a specified criminal offence under the First Schedule of OCHA, or when there is suspicion or reason to believe that the activity is preparatory to the commission of a scam or malicious cyber activity; this includes deepfake images and videos. The Minister said the Government assesses harmful online content holistically and determines the appropriate response based on a range of factors for each case and the legal conditions under the relevant legislation, so that intervention is not determined by any single factor alone. She acknowledged that there are segments of society who may be less confident in their ability to discern between fact and falsehood online, and said public education initiatives are therefore as important as legislation. The Digital Skills for Life framework, developed by the Infocomm Media Development Authority, equips Singaporeans with skills to identify and respond to false and misleading content, while the SG Digital Office runs Gen AI workshops for seniors covering how to stay safe and smart online against AI-generated misinformation.
第 15 届国会 温和质询

监管智能眼镜与AI可穿戴设备以防止隐蔽录制和未经同意的数据收集

Regulating Smart Glasses and AI Wearables to Prevent Covert Recording and Unconsented Data Collection

工人党议员严燕松书面询问数字发展与信息部长:部里是否会出台新规,要求智能眼镜配备可视指示灯以防止在公共场所偷拍;以及如何确保个人使用的AI可穿戴设备不会在未经周围非用户明确同意的情况下大规模采集生物特征或环境数据。部长杨莉明答复:《个人数据保护法》(PDPA)下,组织(包括为商业目的创作内容的个人)须对个人数据的收集和使用负责,拍照、录像或录音适用同样的规则和原则,与设备形态无关。虽然PDPA一般不要求在公共场所拍摄前征得同意,但组织对...

政策信号: 面对AI可穿戴设备带来的新隐私风险,新加坡选择依托技术中立的现有法律而非设备专属立法,监管重心放在数据的后续使用而非采集端硬件。
展开摘要、立场与原文

工人党议员严燕松书面询问数字发展与信息部长:部里是否会出台新规,要求智能眼镜配备可视指示灯以防止在公共场所偷拍;以及如何确保个人使用的AI可穿戴设备不会在未经周围非用户明确同意的情况下大规模采集生物特征或环境数据。部长杨莉明答复:《个人数据保护法》(PDPA)下,组织(包括为商业目的创作内容的个人)须对个人数据的收集和使用负责,拍照、录像或录音适用同样的规则和原则,与设备形态无关。虽然PDPA一般不要求在公共场所拍摄前征得同意,但组织对个人数据的后续使用或披露(包括用于AI功能)必须出于合理目的并遵守其他适用法律。她还指出,未经同意录制、分享或滥用私密内容可能构成《刑法》《防止骚扰法》下的刑事罪行,《网络安全(救济与问责)法》也可追究网络内容创作者对受害者造成伤害的责任。答复未承诺出台针对智能眼镜的设备专属新规。

要点
  • • 严燕松提议要求智能眼镜配备防偷拍可视指示灯
  • • 杨莉明:PDPA规则技术中立,与设备形态无关
  • • 公共场所拍摄一般无需同意,但后续AI用途须合理合法
  • • 刑法、防骚扰法及网络安全法可追责滥用私密内容者
政府立场
政府认为PDPA及刑法等现行法律技术中立、足以覆盖智能眼镜和AI可穿戴设备,暂不出台设备专属的新监管要求。
质询立场
严燕松主张为智能眼镜引入录制可视指示等新规,以防止公共场所偷拍及对旁观者生物特征数据的大规模无授权采集。
"无论设备形态如何,拍照、录像或录音都适用同样的规则和原则。"
原文节选(英文)
Mr Gerald Giam Yean Song asked the Minister for Digital Development and Information (a) whether the Ministry will introduce new regulations for visual indicator requirements for smart glasses to prevent surreptitious recording in public spaces; and (b) how the Ministry ensures that AI-enabled wearable devices used by individuals do not facilitate the mass collection of biometric or environmental data without the explicit consent of non-users in the vicinity. Mrs Josephine Teo replied that under the Personal Data Protection Act (PDPA), organisations, including individuals creating content for commercial purposes, are accountable for how personal data is collected and used, and that the taking of photographs or the making of video or audio recordings is subject to the same rules and principles regardless of the form factor of the device. She noted that while the PDPA generally does not require consent to be sought from persons for photography or recording in public spaces, organisations must ensure that any subsequent use or disclosure of the personal data, including for artificial intelligence features, is for a reasonable purpose and complies with any other applicable laws. Regardless of the capacity in which they are acting, individuals should also note that non-consensual recording, sharing or other misuse of intimate or private content may constitute criminal offences under existing laws, including the Penal Code and the Protection from Harassment Act. Under the Online Safety (Relief and Accountability) Act, creators of online content may also be held accountable for harms caused to victims. The Minister did not announce any new device-specific regulations, such as mandatory recording indicators, for smart glasses or AI wearables.
第 15 届国会 温和质询

评估AI对劳动者竞争力与就业能力的影响及对受影响员工的针对性扶持措施

Assessment of AI Impact on Worker Competitiveness and Employability, and Targeted Support Interventions for Affected Wor...

工人党议员何廷儒就AI对劳动力市场的冲击向人力部长提出两项书面询问:一问AI技能薪资溢价的最新评估及其在行业和资历层面的差异,二问最易被取代的职业评估,以及对文员、后台和入门级知识型岗位员工的针对性干预。部长陈诗龙答复:人力部数据显示,资讯通信、金融服务和专业服务领域的PMET(专业人士、经理、执行员与技师)近年裁员率相对较高,但这反映的是持续重组而非PMET需求萎缩;由于企业AI应用仍处早期,人力部没有AI技能薪资溢价的数据。扶持措施...

政策信号: 政府承认尚无AI对薪资和岗位影响的量化数据,政策重心放在普及AI技能与工具使用,以"提升劳动者"而非"限制企业重组"应对AI冲击。
展开摘要、立场与原文

工人党议员何廷儒就AI对劳动力市场的冲击向人力部长提出两项书面询问:一问AI技能薪资溢价的最新评估及其在行业和资历层面的差异,二问最易被取代的职业评估,以及对文员、后台和入门级知识型岗位员工的针对性干预。部长陈诗龙答复:人力部数据显示,资讯通信、金融服务和专业服务领域的PMET(专业人士、经理、执行员与技师)近年裁员率相对较高,但这反映的是持续重组而非PMET需求萎缩;由于企业AI应用仍处早期,人力部没有AI技能薪资溢价的数据。扶持措施包括:新加坡劳动力局的职业转换计划帮助文员和后台岗位员工转入前景更好的职位,毕业生行业实习计划为未就业应届毕业生提供结构化工作经验;精深技能发展局正在简化AI学习路径;今年晚些时候,参加合格AI培训课程的新加坡人将获得六个月高级版AI工具的免费使用权。政府表示将持续监测劳动力市场并动态调整支持。

要点
  • • 资讯通信、金融与专业服务PMET裁员率较高,官方归因于重组而非需求萎缩
  • • 人力部无AI技能薪资溢价数据,称企业AI应用仍处早期
  • • 职业转换计划、毕业生行业实习计划针对文员后台岗位与应届生
  • • 今年起合格AI课程学员可免费用六个月高级版AI工具
政府立场
政府认为PMET裁员上升反映重组而非需求萎缩,将通过职业转换、毕业生实习、AI学习路径和免费AI工具等组合措施帮助劳动者保持竞争力。
质询立场
何廷儒关注文员、后台和入门级知识型岗位面临的AI取代风险,要求政府给出薪资溢价与高危职业的量化评估及针对性干预。
"今年晚些时候,参加合格AI培训课程的新加坡人将获得六个月高级版AI工具的免费使用权,以培养熟悉度并让他们在真实场景中练习应用这些工具。"
原文节选(英文)
Ms He Ting Ru asked the Minister for Manpower two written questions: first, the Ministry's latest assessment of the salary premium for workers with AI skills in Singapore, including how it varies by sector and seniority, and what further steps will be taken to ensure Singaporean workers acquire these skills and remain competitive; second, the Ministry's latest assessment of the top occupations at risk of displacement, and what targeted interventions are being considered to address employability concerns, especially for workers in clerical, back-office and entry-level knowledge roles. Dr Tan See Leng replied that the Ministry of Manpower's data shows that professionals, managers, executives and technicians (PMETs) in the information and communications, financial services and professional services sectors have seen relatively higher retrenchment incidence in recent years compared to other sectors and occupation groups, and that higher incidence of retrenchment indicates ongoing restructuring rather than a contraction in PMET demand. MOM does not have data on salary premiums for AI skills, as AI adoption by companies in Singapore is still at an early stage, and salaries depend on many factors beyond AI proficiency, including sector, experience and market conditions. On support measures, the Minister cited Workforce Singapore's Career Conversion Programmes, which support individuals, including those in clerical and back-office roles, to pivot into new job roles with good longer-term prospects, and the Graduate Industry Traineeship scheme, which provides fresh graduates who have not found full-time jobs with structured, industry-relevant work experience. SkillsFuture Singapore is making AI learning pathways easier to navigate so working adults can identify courses suited to their proficiency levels and sought by employers. Later this year, Singaporeans who take up eligible AI training courses will receive six months complimentary access to premium versions of AI tools. The Government will continue to monitor labour market trends closely and calibrate support as jobs evolve.
第 15 届国会 实质辩论

不容无就业增长的人工智能(AI)转型(动议提出)

An Artificial Intelligence (AI) Transition with No Jobless Growth (Motion Moved)

全国职工总会(NTUC)秘书长黄志明(惹兰加由议员)5月5日晚提出"AI转型不容无就业增长"国会动议(联署人:Mark Lee、萨克蒂安迪、杨涴凌),要求国会确认AI驱动的增长必须以公平、韧性和人人有机会为锚,新加坡不能出现"无就业增长"。他援引NTUC调查(五分之一受访者把工作保障列为首要担忧、56%的PME自认需提升技能)说明工人焦虑,并提出四项实际举措:建立立足本地的劳动力市场情报与前瞻系统;通过企业培训委员会(CTC,2019年...

政策信号: 劳工运动主动为AI时代重订"经济进步、工人共进"的社会契约,预示新加坡将以三方机制为核心载体,把就业结果与工人保障内嵌进国家AI增长议程而非事后补救。
展开摘要、立场与原文

全国职工总会(NTUC)秘书长黄志明(惹兰加由议员)5月5日晚提出"AI转型不容无就业增长"国会动议(联署人:Mark Lee、萨克蒂安迪、杨涴凌),要求国会确认AI驱动的增长必须以公平、韧性和人人有机会为锚,新加坡不能出现"无就业增长"。他援引NTUC调查(五分之一受访者把工作保障列为首要担忧、56%的PME自认需提升技能)说明工人焦虑,并提出四项实际举措:建立立足本地的劳动力市场情报与前瞻系统;通过企业培训委员会(CTC,2019年以来已设立逾3,800个、惠及逾30万工人)支持企业携工人转型,并提议与全国雇主联合会(SNEF)在三方就业理事会下共同推广;扩大AI-Ready SG等培训路径(目标未来数年逾100万个培训名额);以及让被裁员工有尊严地重新站起,包括裁员前向政府提前通报、把技能创前程求职者援助计划(JSS)覆盖面从月入中位数约5,000新元上调至PME中位收入水平。发言结束后辩论延至次日继续。

要点
  • • 动议四点:承认AI变革力、增长须公平、装备工人与企业、不容无就业增长
  • • NTUC调查:五分之一受访者最忧工作保障,56%的PME认为需提升技能
  • • CTC自2019年已逾3,800个、惠及30万工人,提议与SNEF共同全国推广
  • • 建议JSS覆盖面从月入中位数约5,000新元上调至PME中位收入水平
政府立场
政府已由总理亲自主持全国人工智能理事会、在财政预算案中承诺支持工人和企业适应AI经济,并与劳资双方成立三方就业理事会推动企业转型、职位重设计与工人过渡。
"不是用AI取代工人,而是让AI为所有工种、所有企业的工人服务。"
原文节选(英文)
At 8.29 pm on 5 May 2026, Mr Ng Chee Meng (Jalan Kayu), Secretary-General of NTUC, moved a Motion — also standing in the names of Mr Mark Lee, Mr Saktiandi Supaat and Ms Yeo Wan Ling — that the House recognise the transformative power of AI to drive Singapore's next phase of economic development, anchor AI-enabled growth in fairness, resilience and opportunity for all, equip workers and enterprises to seize new opportunities, and affirm that Singapore must not have jobless growth. Speaking to anxious young graduates, mid-career PMEs and blue-collar workers, he noted that AI is automating routine junior tasks even as it creates new roles; one in five NTUC survey respondents cited job security as their top concern, and 56% of PMEs felt they needed to upskill. He outlined four practical moves. First, build a Singapore-specific system of labour-market intelligence and foresight, combining trade association insights, enterprise data and union sensing. Second, enable enterprises — especially SMEs, which employ about 70% of the workforce — to transform with AI in ways that benefit workers, scaling the Company Training Committee model (more than 3,800 CTCs since 2019, benefiting over 300,000 workers) jointly with SNEF under the new Tripartite Jobs Council; he cited Tan Tock Seng Hospital's AI-enabled PreSAGE fall-risk monitoring system as an example of AI augmenting nursing work. Third, help workers seize new opportunities through e2i's AI Career Coach, NTUC LearningHub skills pathways and AI-Ready SG, which has put more than 4,000 workers through AI training since February, with plans to scale to over one million training places. Fourth, enable displaced workers to bounce back with dignity: advance notification of retrenchments to the Government before the last working day, early deployment of career support through 27 National Career Centres, and expanding SkillsFuture Jobseeker Support coverage from the roughly $5,000 median income closer to PME median income levels. Declaring that "in Singapore, every worker matters", he begged to move; the debate was then adjourned to the next sitting day.
第 15 届国会 温和质询

检讨 2012 年个人数据保护法以应对 AI 生成的推断或衍生数据的使用

Review of Personal Data Protection Act 2012 to Address Use of Inferred or Derived Data Generated by AI

工人党议员陈立峰书面质询数字发展与信息部:是否打算检讨 2012 年《个人数据保护法》(PDPA),以应对推断或衍生数据——包括 AI 生成的行为画像——的使用;若是,检讨将遵循什么原则。部长杨莉明答复:PDPA 下关于可识别个人的数据即属个人数据,机构对其持有或控制的此类数据负有保护义务,也涵盖机构在业务中从个人数据衍生的数据;个人数据保护委员会(PDPC)已发布 AI 推荐与决策系统个人数据使用咨询指引,原则包括仅用于正当商业目的、...

政策信号: 新加坡在 AI 时代的数据保护上延续“现行法 + 软性指引”路线,短期内不会为 AI 推断数据单独修法。
展开摘要、立场与原文

工人党议员陈立峰书面质询数字发展与信息部:是否打算检讨 2012 年《个人数据保护法》(PDPA),以应对推断或衍生数据——包括 AI 生成的行为画像——的使用;若是,检讨将遵循什么原则。部长杨莉明答复:PDPA 下关于可识别个人的数据即属个人数据,机构对其持有或控制的此类数据负有保护义务,也涵盖机构在业务中从个人数据衍生的数据;个人数据保护委员会(PDPC)已发布 AI 推荐与决策系统个人数据使用咨询指引,原则包括仅用于正当商业目的、收集限于所需。答复实质上否定修法迫切性——政府认为现行定义已覆盖衍生数据,辅以软性指引即可,而议员指向的 AI 行为画像新型风险未获专门立法回应。

要点
  • • 政府认定 PDPA 现有定义已涵盖业务过程中衍生的个人数据
  • • PDPC 已发布 AI 推荐与决策系统个人数据使用咨询指引
  • • 指引原则:仅用于正当商业目的、收集限于所需
  • • 未承诺针对 AI 行为画像启动 PDPA 修法检讨
政府立场
政府认为 PDPA 对个人数据的现有定义已覆盖机构在业务中衍生的数据,配合 PDPC 的 AI 咨询指引即可应对,无需立即修法。
质询立场
工人党议员陈立峰认为 AI 生成的推断数据和行为画像构成新型风险,应通过检讨 PDPA 给予专门立法回应。
"这也涵盖机构在业务过程中从此类个人数据衍生出的关于该个人的数据。"
原文节选(英文)
Mr Dennis Tan Lip Fong asked the Minister for Digital Development and Information (a) whether the Ministry intends to review the Personal Data Protection Act 2012 to address the use of inferred or derived data, including behavioural profiles generated by Artificial Intelligence systems; and (b) if so, what principles will guide such a review. Mrs Josephine Teo replied that under the Personal Data Protection Act, data about an identifiable individual is considered personal data, and an organisation has obligations to safeguard such data in its possession or control. She stated that this also covers data about the individual that an organisation derives in the course of business from such personal data — meaning that, in the Government's reading, inferred and derived data already fall within the statute's existing protections rather than sitting in a legislative gap. The Minister further pointed to the Advisory Guidelines on the Use of Personal Data in Artificial Intelligence (AI) Recommendation and Decision Systems published by the Personal Data Protection Commission. These guidelines set out principles to guide organisations and consumers on the responsible collection and use of personal data in AI systems, such as using data only for legitimate business purposes and limiting data collection to what is needed. The answer did not commit to a review of the Act itself, nor did it set out principles for any future review, resting instead on the position that the current statutory definition of personal data, supplemented by PDPC's advisory guidelines, already addresses AI-generated inferred and derived data including behavioural profiles.
第 15 届国会 信息发布

AI 普及对初级律师培养管道的冲击及一年制执业培训框架能否弥补成长断层

Impact of AI Adoption on Junior Lawyer Training Pipelines and Addressing Developmental Gaps Through One-year Practice Tr...

工人党非选区议员刘吴扬(Low Wu Yang Andre)书面询问律政部长:律所大规模采用 AI 后,初级律师赖以磨练专业判断力的研究、起草等常规工作量是否会萎缩,政府是否评估过这一风险;以及新修订律师资格制度下的一年制执业培训框架是否专门针对此风险设计。律政部长唐振辉(Edwin Tong)未作实质答复,表示律政部将把这一问题与其他同主题国会问题合并,在下次适当时机以口头答复方式统一回应,并指引参阅 2026 年 5 月 6 日《官...

政策信号: 法律行业 AI 冲击已进入国会议程,政府选择以合并口头答复统筹回应,显示律政部正将 AI 对律师培养的影响纳入整体性政策叙事而非零散作答。
展开摘要、立场与原文

工人党非选区议员刘吴扬(Low Wu Yang Andre)书面询问律政部长:律所大规模采用 AI 后,初级律师赖以磨练专业判断力的研究、起草等常规工作量是否会萎缩,政府是否评估过这一风险;以及新修订律师资格制度下的一年制执业培训框架是否专门针对此风险设计。律政部长唐振辉(Edwin Tong)未作实质答复,表示律政部将把这一问题与其他同主题国会问题合并,在下次适当时机以口头答复方式统一回应,并指引参阅 2026 年 5 月 6 日《官方报告》中"AI 使用导致律所工作量减少及相关使用指引"一文。背后的张力在于:AI 自动化正侵蚀法律行业传统的学徒式训练路径,而政府对培训管道断层的系统性评估尚未公开。

要点
  • • 工人党议员问 AI 是否削减初级律师赖以成长的研究起草类工作
  • • 并问一年制执业培训框架是否专门应对此风险
  • • 律政部长唐振辉表示将与同主题问题合并口头答复
  • • 实质回应见 2026 年 5 月 6 日合并答复文件
政府立场
律政部不单独书面作答,将把该问题与其他同主题国会问题合并,在下次适当时机统一口头答复。
质询立场
工人党议员刘吴扬质疑 AI 普及正在掏空初级律师的传统训练路径,要求政府说明风险评估及新培训框架的针对性。
"律政部将就此国会问题,连同已就该主题提出的其他国会问题,在下次适当时机提供口头答复。"
原文节选(英文)
Mr Low Wu Yang Andre asked the Minister for Law whether the Ministry has assessed the risk that widespread AI adoption in law firms will reduce the volume of routine work, such as research and drafting, through which junior lawyers have traditionally developed professional judgment; and whether the new one-year practice training framework under the revised admission process is designed to address this risk specifically. The question reflected a concern that as law firms deploy AI tools to automate legal research, document review and first drafts, the apprenticeship pathway through which trainees historically acquired core professional skills could be hollowed out, leaving developmental gaps in the profession's training pipeline. Mr Edwin Tong Chun Fai, the Minister for Law, replied that the Ministry of Law would provide an oral reply to this Parliamentary Question, together with other Parliamentary Questions which had been filed on the same topic, at the next available opportunity. He referred Members to the consolidated response titled "Workload Reduction at Law Firms from AI Use and Guidelines for Such Use", published in the Official Report of 6 May 2026 (Vol 96, Issue 30), in the section for Written Answers to Questions for Oral Answer not Answered by End of Question Time. The substantive assessment of AI's impact on junior lawyer training, and of whether the one-year practice training framework is designed to mitigate that impact, was therefore deferred to the combined oral reply rather than answered individually in this written answer.
第 15 届国会 温和质询

政府安全机构使用的第三方预测分析或数据融合软件算法的独立审计与验证

Independent Audits and Verification of Algorithms Used in Third-party Predictive Analytics or Data-fusion Software Utili...

工人党议员刘吴扬就政府安全机构使用第三方预测分析和数据融合软件提出书面质询:这些机构是否对底层算法进行独立审计?政府能否在依赖此类软件作出行动决策前独立验证其方法论?国家安全统筹部长兼内政部长尚穆根答复称,内政部已建立覆盖 AI 工具开发与使用的治理框架,所有预测分析和数据融合软件——包括第三方供应商提供的底层算法——都须按该框架评估,并针对具体用例实施风险评估与缓解措施;AI 应用部署前后均进行独立评估和审计以确保合规。这一问答触及安...

政策信号: 新加坡将安全领域的第三方 AI 采购纳入部门级治理框架管理,倾向内部审计而非公开外部验证机制。
展开摘要、立场与原文

工人党议员刘吴扬就政府安全机构使用第三方预测分析和数据融合软件提出书面质询:这些机构是否对底层算法进行独立审计?政府能否在依赖此类软件作出行动决策前独立验证其方法论?国家安全统筹部长兼内政部长尚穆根答复称,内政部已建立覆盖 AI 工具开发与使用的治理框架,所有预测分析和数据融合软件——包括第三方供应商提供的底层算法——都须按该框架评估,并针对具体用例实施风险评估与缓解措施;AI 应用部署前后均进行独立评估和审计以确保合规。这一问答触及安全机构算法决策的问责难题:反对党推动外部可验证性,政府则强调内部治理框架已足够覆盖,但未披露审计主体与方法细节。

要点
  • • 内政部设有覆盖 AI 工具开发与使用的治理框架
  • • 第三方供应商的底层算法同样须按框架评估
  • • 按具体用例实施风险评估与缓解措施
  • • AI 应用部署前后均进行独立评估和审计
政府立场
政府认为内政部现有 AI 治理框架加上部署前后的独立评估与审计,足以管控第三方算法在安全行动中的使用风险。
质询立场
工人党议员刘吴扬质疑政府能否在依赖第三方算法作出行动决策前独立验证其方法论,推动更强的外部审计问责。
"在 AI 应用部署前后都会进行独立评估和审计,以确保合规。"
原文节选(英文)
Mr Low Wu Yang Andre asked the Coordinating Minister for National Security and Minister for Home Affairs two related questions about the algorithmic tools used by Government security agencies: first, whether agencies that use third-party predictive analytics or data-fusion software conduct independent audits of the underlying algorithms; and second, whether the Government is satisfied that it can independently verify the methodology of such software before relying on it for operational decisions. Mr K Shanmugam replied that the Ministry of Home Affairs (MHA) has an AI governance framework on the development and use of AI tools. All predictive analytics and data-fusion software used by the ministry, including underlying algorithms supplied by third-party vendors, are evaluated on this framework, so that externally procured algorithms are subject to the same governance treatment as internally developed tools. Beyond the framework-level evaluation, risk assessment and mitigation measures are applied to manage risks arising in specific use cases. Finally, independent assessments and audits are performed both before and after deployment of each AI application to ensure compliance with the framework. The Minister's answer did not name the bodies that conduct these independent assessments, nor describe the audit methodology, but confirmed that verification is built into both the pre-deployment evaluation stage and the post-deployment operation stage of MHA's use of predictive analytics and data-fusion software.
第 15 届国会 信息发布

评估当前网络安全战备能否应对不断演变的威胁并兼顾行动安全

Assessing Adequacy of Current Cybersecurity Readiness against Evolving Threats while Ensuring Operational Security

议员沙拉尔·塔哈以地缘政治紧张、网络行动日益成为混合冲突手段为背景,书面质询政府是否评估新加坡网络威胁暴露已上升,以及如何在不损害行动安全的前提下保护关键信息基础设施、政府系统、企业和个人抵御包括 AI 攻击在内的演变威胁。数字发展与信息部长杨莉明答复:新加坡作为金融枢纽和数字经济体是高价值目标;关键系统受《网络安全法》更高标准约束;网络安全局(CSA)将更新标准,并向关键系统业主提供专有威胁检测系统以应对高级威胁行为者和 AI 赋能威...

政策信号: AI 赋能攻击已被正式纳入新加坡国家网络防御的威胁模型,监管重心从关键基础设施向政府供应链和消费级设备逐层下沉。
展开摘要、立场与原文

议员沙拉尔·塔哈以地缘政治紧张、网络行动日益成为混合冲突手段为背景,书面质询政府是否评估新加坡网络威胁暴露已上升,以及如何在不损害行动安全的前提下保护关键信息基础设施、政府系统、企业和个人抵御包括 AI 攻击在内的演变威胁。数字发展与信息部长杨莉明答复:新加坡作为金融枢纽和数字经济体是高价值目标;关键系统受《网络安全法》更高标准约束;网络安全局(CSA)将更新标准,并向关键系统业主提供专有威胁检测系统以应对高级威胁行为者和 AI 赋能威胁;GovTech 将要求管理关键系统的政府供应商满足 Cyber Trust Mark;家用路由器强制标准从标签计划一级升至二级,并拟扩展至 IP 摄像头。政府承认即使防御最佳仍须对 AI 网络威胁保持警惕。

要点
  • • 关键系统业主将获政府专有威胁检测系统,应对 AI 赋能威胁
  • • 管理关键系统的政府供应商须满足 Cyber Trust Mark 要求
  • • 家用路由器强制标准从标签计划一级升至二级,拟扩展至 IP 摄像头
  • • CSA 的 CISO-as-a-Service 为中小企业提供网络安全顾问
政府立场
政府认为新加坡保持着稳健且自适应的网络安全态势,但承认面对包括 AI 赋能攻击在内的演变威胁必须持续升级标准、检测能力和供应商义务。
"然而,即使拥有最好的防御,我们也必须对不断演变的威胁——包括 AI 赋能的网络威胁——保持警觉和戒备。"
原文节选(英文)
Mr Sharael Taha asked the Minister for Digital Development and Information, in light of rising geopolitical tensions and the increasing use of cyber operations as part of hybrid conflict, whether the Government assesses that Singapore's cyber threat exposure has heightened, and how it assesses Singapore's overall cybersecurity readiness in safeguarding critical information infrastructure, Government systems, businesses and individual residents against evolving threats, including AI-enabled attacks, without compromising operational security. Mrs Josephine Teo replied that Singapore's position as a major financial hub and digital economy makes it an attractive target for malicious actors, with the Cyber Security Agency of Singapore (CSA) regularly updating the public through SingCERT advisories and the Singapore Cyber Landscape publication. Critical systems are held to higher cybersecurity standards and obligations under the Cybersecurity Act, while capability investments such as CSA's Cybersecurity Development Programme strengthen the talent pipeline and national exercises like Exercise Cyber Star enhance operational readiness across public and private sectors. As threats evolve, CSA will review and update cybersecurity standards and obligations, and the Government will help owners of critical systems detect advanced threat actors and AI-enabled threats, including by equipping them with proprietary threat detection systems. For Government systems, GovTech will introduce more stringent cybersecurity and data protection obligations for vendors, such as requiring those managing critical systems and sensitive Government data to meet Cyber Trust Mark requirements. For businesses, CSA's CISO-as-a-Service programme gives small and medium enterprises access to cybersecurity consultants. For citizens, home routers must currently meet Cyber Labelling Scheme Level 1 requirements, which will be raised to Level 2, and similar standards are being explored for IP cameras. The Minister concluded that Singapore maintains a robust and adaptive posture but must remain vigilant against evolving threats, including AI-enabled cyber threats.
第 15 届国会 温和质询

AI 相关技能创前程课程的成效与效果

Outcomes and Effectiveness of AI-related SkillsFuture Programmes

万里扎议员书面质询教育部是否追踪 AI 相关技能创前程课程学员的两项硬指标:结业六个月内是否进入 AI 相关岗位,以及结业 12 个月内工资中位数变化。教育部长李智陞答复:对安置型课程如职业转换计划(SCTP),精深技能发展局(SSG)追踪安置率——2022 年 6 月至 2025 年 3 月完成 ICT 相关 SCTP 的 8,000 名学员中,44% 在六个月内找到新岗位(截至 2025 年 9 月 30 日);但未追踪工资结果,理...

政策信号: 新加坡的 AI 技能再培训体系在产出问责上仍偏软性指标,就业转化与工资增益的硬数据缺口正成为国会关注点。
展开摘要、立场与原文

万里扎议员书面质询教育部是否追踪 AI 相关技能创前程课程学员的两项硬指标:结业六个月内是否进入 AI 相关岗位,以及结业 12 个月内工资中位数变化。教育部长李智陞答复:对安置型课程如职业转换计划(SCTP),精深技能发展局(SSG)追踪安置率——2022 年 6 月至 2025 年 3 月完成 ICT 相关 SCTP 的 8,000 名学员中,44% 在六个月内找到新岗位(截至 2025 年 9 月 30 日);但未追踪工资结果,理由是影响因素多元。短期课程一般不追踪安置结果,仅靠 TRAQOM 学员调查收集质量与自评成效反馈,公开于 MySkillsFuture。张力在于:政府对 AI 技能培训投入巨大,效果衡量却以主观反馈为主,缺乏议员要求的就业与工资硬数据。

要点
  • • 8,000 名 ICT 相关 SCTP 结业学员中 44% 六个月内找到新岗位
  • • SSG 不追踪 SCTP 学员工资结果,理由是影响因素多元
  • • 短期课程一般不追踪就业安置结果
  • • TRAQOM 调查收集学员反馈并公开于 MySkillsFuture 平台
政府立场
政府认为以安置率加学员反馈调查的组合已足以衡量 AI 相关培训成效,工资追踪因干扰因素过多而不可行。
"在 2022 年 6 月至 2025 年 3 月间完成 ICT 相关职业转换计划的 8,000 名学员中,44% 在结业六个月内找到了新岗位或就业。"
原文节选(英文)
Dr Wan Rizal asked the Minister for Education, regarding participants in AI-related SkillsFuture programmes, whether the Ministry tracks (i) employment in AI-related roles within six months of completion and (ii) median wage change within 12 months of completion. Mr Desmond Lee replied that the Ministry of Education and SkillsFuture Singapore (SSG) adopt a variety of quality and outcome metrics to assess the effectiveness of SkillsFuture-funded Continuing Education and Training (CET) programmes, including those that equip learners with AI-related skills. For placement programmes, such as the SkillsFuture Career Transition Programmes (SCTPs), SSG tracks placement rates: of the 8,000 learners who completed their ICT-related SCTP between June 2022 and March 2025, 44% found new roles or employment within six months of completion, as of 30 September 2025. The Ministry has not tracked the wage outcomes of these SCTP learners, as multiple factors can affect their wages. SSG generally does not track the placement outcomes of shorter courses, which are intended to equip learners with up-to-date skills so they can remain relevant and perform better in their current roles rather than transition to new jobs. Beyond placement outcomes, SSG tracks learner feedback through the Training Quality and Outcomes Measurement (TRAQOM) survey, which covers programme quality and perceived outcomes, including whether the training has helped individuals become more effective at their jobs, improve their work performance or take on enhanced responsibilities. The TRAQOM ratings of each course, including AI-related courses, are published on the MySkillsFuture portal to guide learners in selecting suitable training programmes.
第 15 届国会 温和质询

能源危机、AI 数据中心需求与就业前景的冲击

Energy Crisis, AI Data-Centre Demand and the Impact on Hiring Prospects

Yio Chu Kang 议员 Yip Hon Weng 与 Bukit Panjang 议员 Liang Eng Hwa 询问中东冲突引发的能源危机如何冲击招聘前景。人力部长 Tan See Leng 称劳动市场暂时稳健,但企业转趋谨慎——未来三个月有意招聘的企业比例从 2 月的 54.6% 跌至 3 月的 44.6%,4 月初现企稳。被问到长期结构性影响时,他指危机会加速既有结构转型(供应链多元化、数字化),并直接点出 AI 的能源...

政策信号: AI 算力被正式纳入国家能源与就业战略:政府把 AI 数据中心的能源需求视为产业转型与人力规划的核心变量,而非边缘议题
展开摘要、立场与原文

Yio Chu Kang 议员 Yip Hon Weng 与 Bukit Panjang 议员 Liang Eng Hwa 询问中东冲突引发的能源危机如何冲击招聘前景。人力部长 Tan See Leng 称劳动市场暂时稳健,但企业转趋谨慎——未来三个月有意招聘的企业比例从 2 月的 54.6% 跌至 3 月的 44.6%,4 月初现企稳。被问到长期结构性影响时,他指危机会加速既有结构转型(供应链多元化、数字化),并直接点出 AI 的能源含义:"随着 AI 普及,驱动数据中心和高算力需求将需要更多能源,能源已成为新货币。"Choo Pei Ling 追问劳动力规划是否正适应贸易碎片化、技术变革与产业重组叠加的长期不确定,部长以"斩钉截铁的肯定"回应,并提及连续两天合计逾 12 小时(其中谈 AI 七小时)的相关辩论。

要点
  • • 招聘意向企业比例从 2 月 54.6% 跌至 3 月 44.6%,4 月初企稳
  • • AI 普及推高数据中心算力的能源需求,"能源是新货币"
  • • 危机加速供应链多元化、数字化与 AI 采纳等既有结构转型
  • • 人力规划明确调整以应对技术变革与产业重组叠加的长期不确定
政府立场
劳动市场暂稳但企业谨慎;视危机为加速 AI 采纳与能源转型的结构性催化,并据此重塑长期人力规划,能源被定位为支撑 AI 算力的"新货币"
"And with the move towards a pervasive adoption of AI, there will be a need for even more energy to drive the data centres and the high compute requirements of these AI data centres. Energy is really the new currency and it is existential for all of us."
原文节选(英文)
Mr Yip Hon Weng (Yio Chu Kang) and Mr Liang Eng Hwa (Bukit Panjang) asked the Minister for Manpower to assess how the energy crisis triggered by the Middle East conflict was affecting hiring prospects — for graduates and mid-career entrants, on entry-level job creation and job-search duration, the sectors most affected, and what further measures were being considered. The Minister for Manpower, Dr Tan See Leng, addressing both questions together, said energy-intensive and outward-oriented sectors saw the most impact, compounding pressures on export-oriented sectors already hit by trade fragmentation. While the labour market remained resilient for now, firms had grown more cautious: the proportion intending to hire in the next three months fell from 54.6% in February 2026 to 44.6% in March, with early signs of stabilisation in April. Jobseekers could tap Workforce Singapore and SkillsFuture Singapore services. Asked by Mr Yip whether firms were shifting hiring preferences between local and foreign workers, Dr Tan said no such shift had been seen and that resident employment continued to grow in Q1 2026. On Mr Liang's question about permanent structural changes and new opportunities, Dr Tan said it was premature to conclude, but the crisis would likely accelerate shifts already underway — supply-chain diversification and digitalisation — noting Parliament had spent seven hours the previous day on AI adoption, the energy transition and energy resilience. With pervasive AI adoption, he said, even more energy would be needed to drive data centres and their high compute requirements: "energy is really the new currency." Dr Choo Pei Ling (Chua Chu Kang) asked whether workforce planning was being adapted for persistent structural uncertainty from overlapping trade fragmentation, technological change and sectoral restructuring; Dr Tan replied "a resounding yes," referencing the Economic Strategy Review and over 12 hours of recent debate on preparing and investing in workers.
第 15 届国会 信息发布

SIM 卡购买、再注册模式与 GSM 网关滥用的更严管控与实时异常检测

Tighter Controls and Real-time Anomaly Detection for SIM Card Purchases, Re-registration Patterns and GSM Gateway Misuse

赖添来询问数字发展与新闻部是否会对批量 SIM 卡购买、SIM 再注册模式及 GSM 网关滥用引入更严管控与实时异常检测,因这些环节让海外诈骗电话伪装成本地来电。MDDI 高级政务部长陈杰豪回应已布置上游管控:自 2025 年 10 月 1 日起诈骗钱骡被禁开新移动号码,2026 年 2 月 28 日起每人跨所有电信商上限 10 张后付费 SIM(外加既有 3 张预付费上限);2025 年警方瓦解逾 10.5 万条涉诈移动线路。IMDA...

政策信号: AI 与数据分析成为电信层反诈拦截的核心工具,但侦测阈值与细节以行动安全为由保密
展开摘要、立场与原文

赖添来询问数字发展与新闻部是否会对批量 SIM 卡购买、SIM 再注册模式及 GSM 网关滥用引入更严管控与实时异常检测,因这些环节让海外诈骗电话伪装成本地来电。MDDI 高级政务部长陈杰豪回应已布置上游管控:自 2025 年 10 月 1 日起诈骗钱骡被禁开新移动号码,2026 年 2 月 28 日起每人跨所有电信商上限 10 张后付费 SIM(外加既有 3 张预付费上限);2025 年警方瓦解逾 10.5 万条涉诈移动线路。IMDA、政府科技局 (GovTech) 与警方用数据分析侦测可疑购买与注册模式。赖添来追问是否有实时监测、阈值由谁设定。陈杰豪强调这是跨多机构加私营部门的全社会合力,引用"道高一尺,魔高一丈"说明阈值动态调整,并明言出于行动安全不公开细节,但确在运用包括 AI 在内的技术与数据分析。他还披露以电话为首次接触的案件下降、短信为首次接触的案件降约 65%(从 2024 年 1285 起降至 2025 年 450 起),并介绍 IMDA 与 GovTech 联合推出的 SIMCardHowMany 工具供公众自查名下后付费 SIM 数量。

要点
  • • IMDA/GovTech/警方用数据分析侦测可疑 SIM 购买与再注册模式
  • • 部长明言反诈侦测运用包括 AI 在内的技术,但出于行动安全不公开细节
  • • 后付费 SIM 每人上限 10 张(2026-02-28 起),2025 年瓦解逾 10.5 万条涉诈线路
  • • SMS 为首次接触的诈骗案降约 65%;SIMCardHowMany 工具供公众自查名下 SIM
政府立场
以多机构加私营部门合力上游布控,运用 AI 与数据分析动态侦测,阈值跨机构共定且因应手法演变持续调整
"But certainly, we are using technology, data analytics, including leveraging technologies like AI."
原文节选(英文)
Mr Victor Lye (Ang Mo Kio) asked the Minister for Digital Development and Information whether the Government would consider tighter controls and real-time anomaly detection for bulk SIM card purchases, SIM re-registration patterns and GSM gateway misuse, given their role in enabling overseas scam calls to appear local. The Senior Minister of State, Mr Tan Kiat How, replying on the Minister's behalf, said the Government has put in place upstream controls: IMDA has worked with the Singapore Police Force to tighten SIM registration, barring scam mules from new mobile lines since 1 October 2025 and capping each person at 10 postpaid SIMs across all telcos from 28 February 2026, on top of the existing three-prepaid cap. In 2025 the Police disrupted more than 105,000 scam-related mobile lines, and IMDA, GovTech and the SPF use data analytics to detect suspicious purchase and registration patterns. GSM gateways are regulated by IMDA, with devices of five or more SIM slots needing prior approval for import since 1 February 2025, and mules supplying SIMs facing penalties. In supplementary questions, Mr Lye asked whether there is real-time, cross-agency monitoring of unusual data usage and GSM gateways, and who mandates the monitoring thresholds and data criteria. Mr Tan described a multi-agency effort spanning MDDI, MHA, IMDA, GovTech, SPF, ICA and MAS, working with telcos, banks and e-commerce platforms, and said thresholds are reviewed jointly and refined dynamically. Citing the saying (a Chinese proverb about defenders and adversaries perpetually escalating against one another) and operational-security considerations, he declined to share operational details but confirmed the use of technology and data analytics, including AI. He noted that phone-as-first-contact cases have decreased and SMS-as-first-contact cases fell about 65% (from 1,285 in 2024 to 450 in 2025), and pointed to SIMCardHowMany, a tool jointly developed by IMDA and GovTech that lets the public check how many postpaid SIMs are registered under their names and report any discrepancies to their telcos.
第 15 届国会 温和质询

数据中心、城市扩张与气候变率推升需求下对新加坡水安全的审视

Review of Singapore's Water Security Due to Rising Demands from Data Centres, Urban Expansion and Climate Variability

工人党何廷儒询问可持续发展与环境部,在数据中心、城市扩张与气候变率推升用水需求下,新加坡水资源不安全风险的最新评估(尤其 2061 年柔佛供水协议到期前),以及如何加速措施让"四大水喉"满足未来需求而不挤压粮食生产或家庭供应。部长傅海燕回应称公用事业局 (PUB) 持续超前规划并投资水务基建,将经济与人口增长及气候变化纳入考量,并通过海水淡化与新生水等抗气候来源多元化供水。何廷儒追问两点:一是世界银行已指水将从"背景资源"变为"硬约束"...

政策信号: 数据中心算力需求被纳入国家水安全风险账本,但政府仍以既有"四大水喉"框架与降本技术应对,未承诺新增专项措施
展开摘要、立场与原文

工人党何廷儒询问可持续发展与环境部,在数据中心、城市扩张与气候变率推升用水需求下,新加坡水资源不安全风险的最新评估(尤其 2061 年柔佛供水协议到期前),以及如何加速措施让"四大水喉"满足未来需求而不挤压粮食生产或家庭供应。部长傅海燕回应称公用事业局 (PUB) 持续超前规划并投资水务基建,将经济与人口增长及气候变化纳入考量,并通过海水淡化与新生水等抗气候来源多元化供水。何廷儒追问两点:一是世界银行已指水将从"背景资源"变为"硬约束"、水危机或致 2050 年损失高达 6% GDP,包含数据库建设在内的需求上升这一系统性风险如何纳入财政与基建前瞻估算;二是中东冲突推高能源价格,对高度依赖能源的海水淡化成本有何影响、如何应对长期能源价格波动。傅海燕重申已将气候变化纳入长期水务规划,能源成本一向是关注的驱动因素,PUB 持续以新技术降低含能源在内的产水成本,中东事件虽加剧成本压力但不改变长期降本方向。

要点
  • • 数据中心(算力基建)推升用水需求被列为水安全的系统性压力之一
  • • 部长以"四大水喉"+海水淡化/新生水等抗气候来源回应,强调超前规划
  • • 提问方援引世界银行"水成硬约束"与 2050 年或损 6% GDP 框定系统性风险
  • • 海水淡化高度依赖能源,中东冲突加剧成本压力但不改长期降本方向
政府立场
坚持以"四大水喉"超前规划与抗气候水源保障水安全,并持续以新技术压低含能源在内的产水成本
质询立场
主张把数据中心等推升的需求作为系统性风险显式纳入财政与基建前瞻估算,并要求量化中东能源波动对海水淡化成本的冲击
"The World Bank has recently stated that the water will shift from a background resource to a binding constraint."
原文节选(英文)
Ms He Ting Ru (Sengkang) asked the Minister for Sustainability and the Environment, with rising water demands from data centres, urban expansion and climate variability affecting local reservoirs, for the latest assessment of Singapore's water-insecurity risks — particularly ahead of the 2061 Johor Water Agreement expiry — and what accelerated measures are planned to ensure the Four National Taps meet future needs without compromising food production or household supply. The Minister, Ms Grace Fu Hai Yien, replied that PUB continues to plan ahead and invest in water infrastructure to meet future demand, taking into account economic and population growth and the impact of climate change, and has diversified supply through the Four National Taps, including desalinated water and NEWater as weather-resilient sources. In supplementary questions, Ms He raised two points. First, noting that water scarcity could carry implications of up to 6% of GDP by 2050 and that the World Bank has said water will shift from a background resource to a binding constraint, she asked how this systemic-risk framing — including plans to build databanks and climate-driven demand increases — is being incorporated into fiscal and forward estimates for public-infrastructure planning. Second, given that the Middle East conflict has driven up energy costs and made oil and natural gas highly volatile, and that much of Singapore's supply comes from energy-intensive sources such as desalination, she asked about the impact on desalinated-water costs and how Singapore is responding to prolonged energy-price volatility in its desalination cost modelling. Ms Fu replied that her answer had addressed the climate-change dimension, that Singapore plans ahead and incorporates climate-change impact in long-term water-infrastructure planning, and that energy cost is a driver PUB has always attended to — setting energy efficiency as a goal and continually searching for new technology to reduce the overall cost of producing water. She added that recent Middle East events have exacerbated this, but do not change the long-term focus on bringing down water-production costs, including energy.
第 15 届国会 温和质询

针对传播 AI 伪造真人淫秽图像的报警情况

Police Reports on the Circulation of AI-generated Fake Obscene Images of Real Persons

工人党 Sylvia Lim 询问内政部:2025 年收到多少宗关于传播 AI 伪造真人淫秽图像的报警、其中多少涉及教育机构内"同学对同学"、以及受害者能期待怎样的后续处理。代表内政部答覆的高级政务部长 Sim Ann 坦言:警方并不单独统计淫秽物品案件中有多少涉及 AI 伪造真人图像。警方接报后会照常侦查——讯问被告与证人、检视数字鉴证证据;涉及未成年受害者会通知家长/监护人,并向需要心理支持的受害者提供受害者关怀服务;若伪造图像在网...

政策信号: AI 深伪性影像执法:无统计专项、无固定法律路径,依个案事实裁量
展开摘要、立场与原文

工人党 Sylvia Lim 询问内政部:2025 年收到多少宗关于传播 AI 伪造真人淫秽图像的报警、其中多少涉及教育机构内"同学对同学"、以及受害者能期待怎样的后续处理。代表内政部答覆的高级政务部长 Sim Ann 坦言:警方并不单独统计淫秽物品案件中有多少涉及 AI 伪造真人图像。警方接报后会照常侦查——讯问被告与证人、检视数字鉴证证据;涉及未成年受害者会通知家长/监护人,并向需要心理支持的受害者提供受害者关怀服务;若伪造图像在网上流传,警方可依据《在线刑事危害法令》(OCHA) 向网络服务商发出指令,封锁新加坡用户访问。Lim 追问三点:(1) 受害者反映报警后被警方建议改走《防止骚扰法令》(非逮捕性罪行),这是否警方默认做法;(2) 经议员申诉后警方协助下架图像,这是否常规;(3) 《刑事法典》377BD、377BE 等条文本可作为逮捕性罪行处理,何时会这样定性并以起诉为目标侦办。Sim Ann 答:没有"默认走某条路"一说,一切视案情事实而定,若事实构成《刑事法典》下传播私密图像的罪行,警方就会据此行动;网上流传的图像在侦查中一旦掌握,警方可向服务商发指令封锁。

要点
  • • 警方不单独统计涉 AI 伪造真人淫秽图像的案件数
  • • 可依 OCHA 向网络服务商发指令封锁新加坡用户访问
  • • 走《防止骚扰法令》还是《刑事法典》逮捕性罪行视案情而定
  • • 未成年受害者通知家长并提供受害者关怀服务
政府立场
按案情事实定罪定路径,依 OCHA 封锁 + 刑事法典追责并行
质询立场
质疑警方默认引导走非逮捕性的《防止骚扰法令》,促请按逮捕性罪行起诉
"If the AI-generated obscene images are circulating online, the Police may issue directions under the Online Criminal Harms Act to online service providers to disable Singapore users' access to them."
原文节选(英文)
Ms Sylvia Lim (Aljunied, Workers' Party) asked the Coordinating Minister for National Security and Minister for Home Affairs how many police reports were received in 2025 concerning the circulation of AI-generated fake obscene images of real persons; what proportion involved perpetrators and victims who were fellow students in educational institutions; and what victims or complainants can expect by way of follow-up. Replying on the Minister's behalf, Senior Minister of State Ms Sim Ann said the Police do not track how many of the obscene-materials cases they investigate involve AI-generated images of real persons. On any such report, the Police investigate — interviewing the accused and other witnesses and examining available digital forensic evidence; for young victims they may notify parents or guardians, and they offer victim care services to those needing psychological support. Where AI-generated obscene images circulate online, the Police may issue directions under the Online Criminal Harms Act (OCHA) to online service providers to disable Singapore users' access. Ms Lim raised three supplementaries: whether it is the Police's default to advise victims to proceed under the Protection from Harassment Act as a non-arrestable offence; whether the Police routinely assist with image takedowns, as they had after appeals on victims' behalf; and when the Police would instead classify cases under arrestable Penal Code offences such as ss 377BD and 377BE and investigate with a view to prosecution. Ms Sim Ann replied that there is no default route — it depends on the facts of each case; where the facts warrant action under the Penal Code for circulating intimate images, the Police will take it, and obscene images circulating online can be blocked through directions to service providers once identified.
第 15 届国会 温和质询

小学引入与监测 AI 使用的保障措施与路线图

Safeguards and Roadmap for Introducing and Monitoring AI Use by Primary School Students

多名议员(Charlene Chen、Kenneth Tiong、David Hoe 等)联合质询教育部从小学引入 AI 的保障与路线图。教长 Desmond Lee 合并答覆四道问题,端出 MOE "四学"框架——学习关于 AI、学会用 AI、用 AI 来学、最重要的是学会超越 AI。分级路线:小一至小三只做 AI 素养(认知 AI 存在),不布置任何需直接使用 AI 的作业;从小四起,学生已具备基础读写与执行功能,可在教师监督下使用...

政策信号: AI 教育策略:早教素养、晚用工具、护栏内建、纵向研究兜底,拒绝一刀切禁用或全盘数字化
展开摘要、立场与原文

多名议员(Charlene Chen、Kenneth Tiong、David Hoe 等)联合质询教育部从小学引入 AI 的保障与路线图。教长 Desmond Lee 合并答覆四道问题,端出 MOE "四学"框架——学习关于 AI、学会用 AI、用 AI 来学、最重要的是学会超越 AI。分级路线:小一至小三只做 AI 素养(认知 AI 存在),不布置任何需直接使用 AI 的作业;从小四起,学生已具备基础读写与执行功能,可在教师监督下使用专为教育设计、内置护栏的 AI 工具(如 SLS 里的写作助手 LEA、数学 LEA),工具被设计成不会直接喂答案、会以"苏格拉底式"提示把跑题学生拉回正轨;小四起强制 10 小时"Code for Fun"(含编程、计算思维、AI 入门),可选两个 5 小时"AI for Fun"模块(生成式 AI、计算机视觉)。数据匿名化、不用于训练外部模型;用商用 AI 工具须确保输入不含个人可识别信息。研究上由 A*STAR 的 SG-LEADS 纵向研究(2027 起采数)追踪儿童 AI 使用对学习与福祉的影响,同时配合学校做短期研究。Kenneth Tiong 以瑞典 Karolinska 研究院"数字工具有害学习"结论与瑞典 2023 年逆转数字化(投 2 亿欧元重回纸质课本)施压,问 MOE 的分歧点何在;Desmond Lee 答:"瑞典是发到五岁、全盘数字化后又全盘退回纸质,我们走的是混合路线"——继续保留实体课本与教师中心教学,把 AI 当工具,且严格区分通用 AI 与专为教育设计的 AI,否则会走错政策方向、干脆完全不用 AI,那才是错误。家长可否退出:若是 SLS 课堂工具属教学一部分不可退出;若是外部引入、需家长同意的工具,未获同意则不让用。Eileen Chong 提"公平悖论"——更弱势、家中缺成人监督的孩子可能更依赖 AI、反被侵蚀认知发展;教长答这是"常青"议题,靠 AI 素养内化 + 家校社区伙伴关系应对。

要点
  • • MOE "四学"框架:学习、使用、用 AI 学、超越 AI
  • • AI 素养小一起教,工具使用小四起且仅限教师监督下的教育专用工具
  • • SLS 工具不喂答案、苏格拉底式提示防认知卸载
  • • 回应瑞典逆转:走混合路线,严格区分通用 AI 与教育 AI
政府立场
分级、混合、教师监督的 AI 素养路线,严格区分通用 AI 与教育专用 AI
质询立场
援引瑞典逆转质疑早引入 AI 的科学依据,担忧认知卸载与弱势学生的公平悖论
"It is far better to start AI literacy young and start getting our kids to use AI for learning in a highly scaffolded teacher supervised and of course, parents supervised way."
原文节选(英文)
Dr Charlene Chen (Tampines), Mr Kenneth Tiong Boon Kiat (Aljunied), Mr David Hoe (Jurong East-Bukit Batok) and others asked the Minister for Education about safeguards, prerequisite skills, the Primary 4 entry point, the implementation roadmap, data transparency, procedural fairness and teacher support for AI use by primary-school pupils. Minister Desmond Lee answered the questions together. He set out the "Four Learns" — learn about AI, learn to use AI, learn with AI, learn beyond AI — and a calibrated, learning-sciences-informed roadmap: AI literacy from Primary 1 with no direct-use work in P1–3; from Primary 4, supervised use of purpose-built educational AI tools with built-in guardrails (the writing assistant LEA and Maths LEA in SLS), which refuse to spoon-feed and redirect off-task pupils Socratically; a mandatory 10-hour Code for Fun programme and optional AI for Fun modules. Pupil data is anonymised and not used to train external models. A*STAR's SG-LEADS longitudinal study will collect data from 2027 to track AI's effect on children's learning and well-being. Dr Chen asked about safeguarding independent thinking, vulnerable and SEN pupils; Mr Tiong pressed on productive struggle, metacognition and Sweden's reversal of digitalisation; Mr Hoe raised teachers asking pupils to use public tools like ChatGPT, age-restriction consent and letting parents try the tools; Ms Eileen Chong Pei Shan (NCMP) raised the equity paradox; Ms Cassandra Lee and Ms Elysa Chen asked about age-based safeguards, the AI "nutrition label" and IP/creator rights. Mr Lee replied that Singapore takes a blended approach — retaining physical textbooks and teacher-centric teaching, treating AI as a tool and strictly distinguishing general-purpose AI from educational AI — and that AI literacy, the home-school partnership and continued calibration are the core safeguards.
第 15 届国会 温和质询

美国判定 Meta 与 Alphabet 存在过失后对青少年社媒访问的监管

Regulating Young People's Access to Social Media after US Negligence Finding against Meta and Alphabet

Vikram Nair 援引美国法院判定 Meta 与 Alphabet 在平台设计上对青少年造成伤害存在过失,问数字发展与新闻部会否监管青少年使用社媒、乃至考虑禁令。代表 MDDI 答覆的政务部长 Rahayu Mahzam 先指此前书面答覆已涵盖,再回应追问:政府正以该判决为参照,相关发现"为不断累积的证据体系再添一笔",伤害是真实的、政府严肃对待——但走的是不同路径:要更有效、更经久、能抵御技术演变的做法。她强调判决揭示的是社媒"...

政策信号: 青少年在线安全:从内容管制转向设计特征监管,AI 伴侣被列为新兴风险,禁令保留为后手
展开摘要、立场与原文

Vikram Nair 援引美国法院判定 Meta 与 Alphabet 在平台设计上对青少年造成伤害存在过失,问数字发展与新闻部会否监管青少年使用社媒、乃至考虑禁令。代表 MDDI 答覆的政务部长 Rahayu Mahzam 先指此前书面答覆已涵盖,再回应追问:政府正以该判决为参照,相关发现"为不断累积的证据体系再添一笔",伤害是真实的、政府严肃对待——但走的是不同路径:要更有效、更经久、能抵御技术演变的做法。她强调判决揭示的是社媒"特定功能与维度"致害,而非平台本身,因此 MDDI 的做法针对具体伤害与产生伤害的具体设计特征、并按用户年龄校准——这比一刀切禁令"更严格、更费力"。新加坡并非从零起步:已有《在线安全行为准则》、对应用引入年龄核验、要求指定社媒服务年度报告,近期已据此对 X 与 TikTok 采取行动。下一步将把年龄核验扩展到指定社媒服务,并从内容延伸到设计特征——陌生人私信、自动播放、其他驱动过度使用的功能,以及"AI 伴侣(AI companions)带来的新兴风险"。政府不排除禁令选项("会不惜一切保护年轻人"),但全面禁令并非全球共识——爱沙尼亚、比利时、纽约州未采用,澳洲也已修法转向更有针对性、聚焦设计特征的做法。

要点
  • • 以美国 Meta/Alphabet 过失判决为参照,认伤害真实
  • • 走针对具体伤害与设计特征、按年龄校准的路线,非一刀切禁令
  • • 已有在线安全准则 + 年龄核验,近期对 X、TikTok 采取行动
  • • 将监管从内容延伸到设计特征,含 "AI 伴侣" 新兴风险
政府立场
不一刀切禁社媒,按年龄校准、靶向致害设计特征(含 AI 伴侣),不排除最终禁令
"Other features that drive excessive use, emerging risks from AI companions and all that. So, we are actually going to break it down and see how we can improve it."
原文节选(英文)
Mr Vikram Nair (Sembawang) asked the Minister for Digital Development and Information, in light of the US Court's finding that Meta and Alphabet were negligent in designing platforms that harmed young people, whether the Government would consider action to regulate young people's access to social media, including a potential ban. Minister of State Ms Rahayu Mahzam, replying for MDDI, noted the matter had been addressed in earlier written replies, then took the supplementary. She said the Government is taking reference from the judgment, whose findings add to a growing body of evidence; the harms are real and taken seriously, but the Government prefers a different, more effective and more durable approach that can withstand technological change. Because the judgment showed that specific features and dimensions of social media — not the platforms wholesale — cause harm, MDDI targets the specific harms and the design features producing them, calibrated by the user's age, which she described as more demanding and more rigorous than a blanket ban. She noted Singapore is not starting from a blank slate: it has the Code of Practice for Online Safety, age assurance for apps and annual reporting by designated social media services, and has acted against X and TikTok. Going forward, it will extend age assurance to designated social media services and look beyond content to design features — direct messaging from strangers, auto-play, other features that drive excessive use, and emerging risks from AI companions. The Government is not foreclosing a ban and will do whatever it takes to protect young people, but observed that a blanket ban is not a globally accepted position, citing Estonia, Belgium, New York State and Australia's move toward a more targeted, design-feature-focused approach.
第 15 届国会 温和质询

应对前沿 AI 模型窃取数据、破坏关键基础设施、利用软件漏洞的风险

Response to Risks from Frontier AI Models with Potential to Steal Data, Disrupt Critical Infrastructure and Exploit Soft...

Saktiandi Supaat 与 Edward Chia 就前沿 AI 模型(点名 Anthropic 的 Mythos、能自主发现并利用软件漏洞)对新加坡金融体系与关键基础设施的威胁提问,并问这是否构成"新类别的系统性金融风险"。数字发展与新闻部高级政务部长陈杰豪代部长答覆,核心立场是把这视为"连续谱上的能力提升"而非"断点式跃迁":政府并未取得 Mythos 访问权(Anthropic 仅向少数伙伴受控预览,无本地银行获权),但...

政策信号: AI 网络威胁应对升至董事会/CEO 层级,从单一模型转向跨部门体系化防御
展开摘要、立场与原文

Saktiandi Supaat 与 Edward Chia 就前沿 AI 模型(点名 Anthropic 的 Mythos、能自主发现并利用软件漏洞)对新加坡金融体系与关键基础设施的威胁提问,并问这是否构成"新类别的系统性金融风险"。数字发展与新闻部高级政务部长陈杰豪代部长答覆,核心立场是把这视为"连续谱上的能力提升"而非"断点式跃迁":政府并未取得 Mythos 访问权(Anthropic 仅向少数伙伴受控预览,无本地银行获权),但 OpenAI GPT-5.5 已有可比的网络能力且更易获得,开源模型几个月内会追上。他举证 AI 已实质改变攻击形态——Google 2025 年报告的 PROMPTFLUX 恶意软件会在攻击中实时调用 AI 改写自身代码以躲避检测;2024 年有犯罪集团用 AI 深度伪造视频通话冒充跨国公司 CFO,骗员工转账 2560 万美元。政府定性这是"对既有系统性风险的放大"而非全新类别。具体动作:MAS 已召集各大金融机构 CEO 推动集体网络韧性行动;CSA 当天向全部 11 个关键信息基础设施(CII)部门的董事会与高层发函,要求据 AI 威胁重审网络风险姿态。陈杰豪给出五点要务(重做风险评估、盘清资产、更快打补丁+持续监控、治理自身 AI 使用、用 AI 做防御),并强调政府正自研 AI 网络防御能力以免受制于单一外部方。

要点
  • • 前沿 AI 模型威胁被定性为"对既有系统性风险的放大",非全新类别
  • • 政府无 Mythos 访问权,但视其为能力连续谱而非断点跃迁
  • • CSA 当天向全部 11 个 CII 部门董事会发函要求重审网络风险姿态
  • • MAS 已召集各大金融机构 CEO 推动集体网络韧性
政府立场
把 AI 网络风险视为既有系统性风险的放大;坚持"基本功优先 + 用 AI 防御",政府自建能力
"With AI, vulnerabilities that once took expert teams weeks to detect manually can now be identified autonomously in hours, sometimes minutes."
原文节选(英文)
Mr Saktiandi Supaat and Mr Edward Chia Bing Hui (Holland-Bukit Timah) asked the Minister for Digital Development and Information about the risks from advanced AI models — including Anthropic's Mythos, reported to autonomously identify and exploit software vulnerabilities — to Singapore's financial system and critical infrastructure, whether such AI-enabled cyber risks constitute a new class of systemic financial risk, and whether existing cybersecurity frameworks remain fit for purpose. Replying for the Minister, Senior Minister of State Mr Tan Kiat How said the Government does not have access to Mythos and does not assume early access to every frontier model; instead it maintains working relationships with AI labs and cybersecurity firms to track capabilities. He stressed the shift is a continuum, not a step change: GPT-5.5 already shows comparable capabilities and open-source models are catching up. He cited PROMPTFLUX malware that consults a live AI model to rewrite its code mid-attack, and a 2024 deepfake CFO fraud that moved US$25.6 million. The Government views this as amplifying existing systemic risk rather than a wholly new category. MAS convened major financial institutions' CEOs on cyber resilience; CSA issued a letter that day to all 11 CII sectors' boards and senior leadership to review their cyber risk posture. Mr Tan set out five priorities — revisit risk assessments, know your assets, patch faster with continuous monitoring, govern your own use of AI (per CSA's "Securing Agentic AI" addendum), and use AI in defence. In supplementaries, Mr Edward Chia, Mr Yip Hon Weng and Mr Louis Chua pressed on international norms, framework adequacy, SME support, talent depth, and direct access to Mythos; Mr Tan reiterated that fundamentals matter, that the Government is building in-house AI cyber-defence tools to share with CII owners, and that he has personally engaged the leadership of all 11 CII sectors.

分析视角

从完整档案中提炼出的解释层。

💡 核心洞察

数据主权成核心议题

重要

随着AI产业发展,数据主权和防止外资垄断成为持续关注焦点,政府强调国际合作与安全防护并重。

新加坡AI经济与数据主权,AI企业收购与本地人才保障

劳动力转型政策深化

重要

政府推动包容性劳动力转型,支持中年员工再培训及技能转移,回应AI替代与互补争议。

AI对中小企业劳动力影响研究,就业转型图及中年员工支持

AI教育多元融合

教育政策强调分龄AI使用指导、脑科学融合及多元教学法,平衡学术严谨与创新能力培养。

分龄生成式AI使用框架,脑科学与自适应AI教育合作,PSLE学术严谨与AI技能平衡

监管机制趋于灵活

AI企业并购采用自愿申报,监管强调创新与竞争平衡,存在监管充分性争议。

Meta收购Manus的监管评估

AI伦理与透明度提升

针对AI生成内容声明及广告规范,政府推动指引更新,兼顾消费者保护与行业创新。

物业代理AI编辑图像声明

医疗AI预防性应用加速

重要

卫生部推出ACE-AI工具预测慢性病风险,强调'AI增强而非AI决定'原则,BRCA基因检测补贴与MediShield Life扩展标志医疗AI进入实质部署阶段。

2026卫生部供给委员会辩论:ACE-AI工具部署、BRCA1/2基因检测补贴

技能培训资金支持不足

AI相关培训课程参与度提升,资金支持及课程定制仍为争议焦点,政府持续优化。

SkillsFuture AI培训参与率及支持,扩展SkillsFuture信用涵盖AI工具订阅

📈 政策演变脉络

点击任意时期,筛选该阶段的辩论记录。

⚡ 反复争议焦点

点击任意争议焦点,筛选相关辩论。

⚖️ 核心政策张力

创新速度 vs 监管安全

鼓励快速创新与市场活力
vs
强调风险控制与公平竞争
当前平衡: 采用灵活监管,鼓励自愿申报

数据开放 vs 数据主权

推动国际数据合作与共享
vs
防范外资垄断与数据流失
当前平衡: 强调数据安全与国际合作并重

AI替代 vs 劳动力转型

AI替代部分岗位提升效率
vs
保障就业,支持再培训转岗
当前平衡: 推动包容性转型与技能提升

教育公平 vs 创新能力培养

注重学术严谨与核心能力
vs
强调探究式与协作式创新学习
当前平衡: 融合多元教学,兼顾公平与创新

资金支持 vs 培训课程定制

增加资金支持扩大培训覆盖
vs
优化课程内容满足多样需求
当前平衡: 持续优化课程与资金配置

🎙️ 关键议员档案

点击任意议员卡片,筛选该议员参与的辩论。

📡 政策信号追踪

点击任意信号,筛选该政策信号首次出现的年份。

继续探索