EU landmark rules for ChatGPT give Singapore the blueprint to rein in AI harms: Experts

The European Union on Aug 31 added the AI chatbot to a list of digital services subjected to the bloc’s toughest rules for online safety.

Published Sep 06, 2026, 05:00 AM

Updated Sep 06, 2026, 05:00 AM

The EU’s strict rules on AI platforms like ChatGPT and social media safety are providing Singapore with evidence to enhance its own online harm regulations, especially for protecting children.

Singapore is already adopting similar measures to the EU, including age verification and risk mitigation on social media.

Experts suggest Singapore will balance innovation with safety, gradually updating safeguards as evidence grows.

SINGAPORE – The European Union’s landmark move in late August to bring artificial intelligence chatbot ChatGPT, online game Roblox and messaging platform Reddit under the bloc’s strictest rules for online safety is expected to give Singapore more evidence to rein in potential harms.

The European rules include adding age checks to prevent minors from being exposed to adult content such as pornography and gambling, and changing how systems recommend content to children to lower their risk of encountering harmful material.

The EU guidelines also spell out that children and teens’ online accounts be kept private by default to avoid them being contacted by strangers.

Dirk Hartung, an assistant professor at Singapore Management University’s Yong Pung How School of Law, said that it becomes easier for Singapore lawmakers to also require similar measures here.

“Companies will have a hard time arguing that they cannot deploy the same tools to protect users in Singapore,” he said.

Experts said that developments in the United States also provide points of reference.

Earlier in August, Meta agreed to a settlement which included adding more safeguards for children and teenagers on Facebook and Instagram, such as limiting their daily use of the social media platforms and restricting children’s use at night.

“The measures being considered or implemented overseas – including age assurance, safer default settings and restrictions on potentially harmful design features – provide useful evidence that can inform Singapore’s own regulatory approach,” said Ben Chester Cheong, a lawyer at Rajah and Tann Singapore and a law lecturer at the Singapore University of Social Sciences.

Noting that many jurisdictions across the globe are also tightening safety measures, Hartung said regulators will step in where addiction, self-harm and other negative consequences for children’s well-being are accepted as part of online platforms’ business model.

Still, Cheong said that Singapore is already moving in a comparable direction as the EU in two areas of online safety.

The EU requires very large online services to assess and mitigate risks arising from their systems, including online harms from how a platform recommends, amplifies or delivers content. Similarly in Singapore, designated social media services must establish systems and processes to reduce exposing users to harmful material.

Measures in the EU to protect children, including age verification, are also moving forward in Singapore for social media.

Beyond EU and US rules, other measures Singapore could consider include holding online platforms to higher levels of transparency and accountability, such as disclosing their funding sources.

Samuel Yuen, managing director of Yuen Law, said that AI can be manipulated to drive fake news.

“Knowing the source of an AI platform’s funding allows us to understand if there are conflicts of interest between a funder and how an AI is used,” he said, such as a biased AI that gives only positive replies about a funder.

Hartung added that making it clear to users that they are interacting with AI chatbots is helpful so they are not tricked into thinking they are engaging with a real person instead of a machine.

But the Singapore Government is unlikely to follow EU and US measures to a tee. Instead, Cheong said that the Republic will likely assess overseas evidence on online harms alongside Singapore’s own experience and adopt measures that are commensurate with the risks here.

Yuen said that this means developing Singapore’s own response, according to its cultural values, economic realities and national interests.

“The key is about finding an approach and framework to cast our own net. But I believe our lawmakers will move at the speed of necessity,” he said.

Agreeing, Cheong said that Singapore can move quickly once a risk is sufficiently established.

He cited how in April, Singapore required app stores to screen and block users under 18 from downloading apps not meant for their age. Also, in June, the Republic’s Online Safety Commission began operations to help victims of online harms seek redress through content takedown orders.

In early 2027, changes are expected to require social media platforms here to enforce safety features through age checks.

While the EU typically takes a stricter stance on public safety and privacy, Goh Kim Huat, a professor at Nanyang Technological University’s Nanyang Business School, said that Singapore tends to take an approach that “calibrates the needs of businesses while safeguarding the interest of individuals”.

Said Cheong: “This balance between trust, safety and innovation has been a consistent feature of Singapore’s approach to digital and AI governance.”

“Singapore’s broader approach is to allow innovation to develop while continually updating safeguards as evidence develops,” he added.

AI/artificial intelligence

8月31日，欧盟将人工智能聊天机器人添加到受欧盟最严格的网络安全规则约束的数字服务列表中。

发布于 2026 年 9 月 6 日上午 5:00

更新于2026年9月6日上午5:00

欧盟对 ChatGPT 等人工智能平台和社交媒体安全的严格规定，为新加坡提供了加强自身网络危害法规（尤其是保护儿童的法规）的证据。

新加坡已经采取了与欧盟类似的措施，包括社交媒体上的年龄验证和风险缓解措施。

专家建议新加坡将在创新与安全之间取得平衡，随着证据的积累逐步更新保障措施。

新加坡——欧盟在8月下旬采取了一项具有里程碑意义的举措，将人工智能聊天机器人ChatGPT、在线游戏Roblox和即时通讯平台Reddit纳入欧盟最严格的网络安全规则之下。预计此举将为新加坡提供更多证据，以遏制潜在的危害。

欧洲的相关规定包括增加年龄检查，以防止未成年人接触色情和赌博等成人内容，并改变系统向儿童推荐内容的方式，以降低他们接触有害内容的风险。

欧盟指南还明确规定，儿童和青少年的在线账户默认应保持私密状态，以避免陌生人与他们联系。

新加坡管理大学杨邦豪法学院助理教授德克·哈通表示，新加坡立法者更容易要求在这里也采取类似的措施。

他说：“企业很难辩称他们无法部署同样的工具来保护新加坡的用户。”

专家表示，美国的发展情况也提供了参考点。

8 月初，Meta 同意了一项和解协议，其中包括为 Facebook 和 Instagram 上的儿童和青少年增加更多保护措施，例如限制他们每天使用这些社交媒体平台的时间，以及限制儿童在夜间使用这些平台。

“海外正在考虑或实施的措施——包括年龄保证、更安全的默认设置以及对潜在有害设计特征的限制——提供了有用的证据，可以为新加坡自身的监管方法提供参考，”新加坡拉惹律师事务所的律师、新加坡社会科学大学的法律讲师张俊杰表示。

哈通指出，全球许多司法管辖区也在加强安全措施，如果网络平台将成瘾、自残和其他对儿童福祉造成负面影响的行为视为其商业模式的一部分，监管机构将会介入。

不过，张先生表示，新加坡在网络安全的两个领域已经朝着与欧盟类似的方向发展。

欧盟要求大型在线服务提供商评估并降低其系统带来的风险，包括平台推荐、放大或传播内容的方式可能造成的网络危害。同样，在新加坡，指定的社交媒体服务提供商必须建立相应的系统和流程，以减少用户接触有害内容的机会。

欧盟为保护儿童而采取的措施，包括年龄验证，也在新加坡的社交媒体领域得到推进。

除了欧盟和美国的规定之外，新加坡还可以考虑采取其他措施，例如要求在线平台提高透明度和问责制，例如披露其资金来源。

元罗律师事务所董事总经理袁先生表示，人工智能可以被操纵来制造虚假新闻。

他说：“了解人工智能平台的资金来源，可以让我们了解资助者和人工智能的使用方式之间是否存在利益冲突”，例如，人工智能可能存在偏见，只对资助者做出正面回应。

哈通补充说，明确告知用户他们正在与人工智能聊天机器人互动是有帮助的，这样他们就不会被欺骗，以为自己是在与真人而不是机器互动。

但新加坡政府不太可能完全照搬欧盟和美国的措施。张先生表示，新加坡可能会结合自身经验，评估海外关于网络危害的证据，并采取与本地风险相称的措施。

袁先生表示，这意味着要根据新加坡的文化价值观、经济现实和国家利益，制定新加坡自己的应对措施。

他说：“关键在于找到一种方法和框架来撒下我们自己的网。但我相信我们的立法者会根据实际需要以最快的速度采取行动。”

张先生表示同意，并指出一旦风险得到充分证实，新加坡就可以迅速采取行动。

他举例说，新加坡在4月份要求应用商店对18岁以下用户进行筛选，并阻止他们下载不适合其年龄的应用。此外，新加坡网络安全委员会在6月份开始运作，帮助网络侵害的受害者通过下架侵权内容来寻求赔偿。

预计到 2027 年初，相关规定将要求本地社交媒体平台通过年龄检查来强制执行安全功能。

尽管欧盟通常在公共安全和隐私方面采取更严格的立场，但南洋理工大学南洋商学院教授吴金发表示，新加坡倾向于采取一种“在保障个人利益的同时，兼顾企业需求”的做法。

张先生表示：“信任、安全和创新之间的这种平衡一直是新加坡在数字和人工智能治理方面一贯的特点。”

他补充说：“新加坡的总体方针是允许创新发展，同时随着证据的发展不断更新保障措施。”

人工智能