口头答复 · 2026-09-09 · 届国会 15
为借助生成式 AI 寻求宗教解答的穆斯林提供指引
为借助生成式 AI 寻求宗教解答的穆斯林提供指引
议员安迪(Mr Saktiandi Supaat)询问代主管回教事务部长:新加坡伊斯兰宗教理事会(MUIS)是否会发布指引,帮助借助生成式 AI 获取宗教解答的穆斯林核实答案。代主管回教事务部长扎吉哈(Mr Zaqy Mohamad)答复:生成式 AI 可补充宗教学习,但不应作为宗教指引的主要来源,复杂问题应请教合格宗教师;MUIS 已通过星期五讲道、公开讲座和 MuslimSG 平台传达这一信息。补充提问中,安迪要求提供核查 AI 答案的渠道,扎吉哈表示 MUIS 正研究升级 MuslimSG 以便公众接触宗教师,部分宗教师已受 AI 培训。议员文礼佳(Dr Wan Rizal)建议设专门单位主动搜寻网上宗教错误信息,扎吉哈认为互联网过于庞大、算法难以监管,重点应是提升民众辨识风险的能力。
为什么重要
MUIS 将升级 MuslimSG 并培训宗教师使用生成式 AI,但未接受设专门单位巡查网上宗教错误信息的建议
关键要点
- • 政府立场:生成式 AI 和网络平台可补充宗教学习,但不应作为宗教教导或指引的主要来源,复杂宗教问题应请教合格宗教师
- • MUIS 已通过星期五讲道、公开讲座和 MuslimSG 平台传达这一信息,并将通过社交媒体和社区伙伴提供实用指引,加强数字宗教素养
- • MUIS 正研究升级 MuslimSG 平台,提高接触宗教师的便利性,未来可能使其更具互动性
- • 部分宗教师已接受培训,以了解 AI 的影响及如何更好地使用 AI;公众可通过宗教师认证计划网站联系宗教师
- • 扎吉哈前一天宣布成立的网络委员会也负责研究网络来源和网络伤害带来的新兴风险
- • 对设专门单位巡查网上宗教错误信息的建议,代部长认为互联网过于庞大、算法难以监管,应着重识别主题并提升民众辨识风险的能力
政府认为生成式 AI 只能作为宗教学习的补充,复杂宗教问题须请教合格宗教师,MUIS 将升级 MuslimSG、培训宗教师并加强数字宗教素养。政府未承诺设立专门单位主动搜寻网上宗教错误信息,认为算法难以监管,重点应放在识别主题和提升民众辨识风险的能力。
提问议员安迪担心寻求宗教指引的人缺乏核实 AI 答案的知识,希望 MUIS 提供可提交答案核查的渠道,并触及不常参加清真寺活动的人群;议员文礼佳认为现有做法过于被动防守,建议设专门单位主动巡查。
新加坡应对生成式 AI 带来的宗教错误信息,倾向于依靠宗教机构的指引、宗教师培训和公众素养,而不是监控内容来源,这与其在网络伤害问题上强调给民众“打预防针”的整体思路一致。
“生成式 AI 也不是来自网站;它来自从不同网站汇集内容的算法,因此很难监管算法。”
参与人员 (3)
完整译文(中文)
Hansard 原始记录 · 2026-09-26
12 号,安迪先生询问代主管回教事务部长:新加坡伊斯兰宗教理事会(MUIS)是否会发布指引,帮助使用生成式人工智能及其他网络来源获取宗教解答的穆斯林,对照权威来源核实这些答案,并了解何时、何处应请教合格的宗教师(asatizah)。
代主管回教事务部长(扎吉哈先生):议长先生,生成式人工智能(GenAI)和网络平台能让信息与知识更易获取,从而对宗教学习起到补充作用。然而,它们不应被当作宗教教导或指引的主要来源,尤其是在需要学术诠释和语境理解的复杂宗教问题上。在这些情况下,应当请教合格的宗教师或其他值得信赖的知识来源。
新加坡伊斯兰宗教理事会(Majlis Ugama Islam Singapura,简称 MUIS)一直通过星期五讲道、公开讲座和 MuslimSG 平台强化这些信息。MUIS 将继续努力,通过社交媒体和我们的社区伙伴网络提供实用指引,以加强数字宗教素养。
议长:安迪先生。
安迪先生(碧山-大巴窑):谢谢议长先生。我有两个补充问题。感谢代部长的答复。寻求宗教指引的人即使获得了参考资料,也未必具备核实 AI 生成答案的知识。因此,我的第一个补充问题是:MUIS 能否加强现有渠道,让公众可以提交这类答案供核查,例如由合格的宗教师核查,并及时得到回复,特别是涉及家庭或个人决定等与宗教问题相关的重要决定时?
第二个补充问题是:正如您所提到的,一些穆斯林或马来族/穆斯林的宗教信息大多来自网上,其中包括可能不常参加清真寺活动或宗教课程的人,MUIS 将如何确保指引能够触及他们?例如,MUIS 今后是否会提供实际例子,帮助他们识别哪些 AI 生成的答案需要核查、哪些需要由合格的宗教师核查?
扎吉哈先生:谢谢议长先生。这些确实是合理的问题,也是令人关注的问题,不仅穆斯林社群关注,我认为更广泛的社会也关注,即如今人们如何消费来自世界各地的内容。影响也来自全球各地。
因此,我认为问题实际上在于,我们如何让不仅是信众(Jemaah),还有整个穆斯林群体(ummah),乃至更广泛的国民,更好地认识到需要提高警觉、能够核查信息,无论是宗教内容还是其他内容。我认为,归根结底,这是我们作为一个国家需要努力的事情。
有两点:第一,关于议员问我们能否让平台更易使用,答案是肯定的。我们在提高信息的可及性、例如接触宗教师的便利性方面,确实还有改进空间。这是 MUIS 正在研究的一个领域,即升级 MuslimSG 平台。
我们相当密切地关注生成式 AI。因此,我们正在教育甚至提升许多宗教师的能力,让他们更熟悉生成式 AI 和新的数字工具。其中一些人已接受培训,以更好地了解 AI 的影响或如何更好地使用 AI,因为归根结底,他们必须作出解释。
第二,我认为更广泛的问题是,我们如何让穆斯林社群更好地利用各种平台,包括实体和网络平台。因为我认为,我们看待宗教教育必须采取全面的方式,兼顾实体领域和数字领域。如今宗教师是可以接触到的,大家可以通过宗教师认证计划(Asatizah Recognition Scheme)网站联系他们。当然,如果有任何疑虑或需要任何宗教指引,也可以通过 MUIS 联系我们。今后在升级 MuslimSG 时,我们或许可以让它更具互动性,也更便于社群使用。
我认为,归根结底,这是我们非常清楚的一个领域,尤其是在设法接触可能受到影响的年轻人方面,特别是存在激进化风险,甚至所谓“ajaran sesat”即错误教义的风险。
在宗教师群体中,我认为关键在于能够把当今许多人从海外教士或穆夫提那里获得、接收或受其影响的一些复杂问题放到具体语境中理解。我认为这些都是重要方面,我们需要能够帮助社群把许多这类教义重新放回新加坡的语境中理解。
因此,有许多层面需要努力,我们今天也正在努力。但我向议员保证,这是我们正在关注的领域。我昨天宣布成立的网络委员会,也负责研究来自网络来源和网络伤害的一些新兴风险。我们确实在关注这个问题,既着眼于现状,也主动审视新兴风险。
议长:文礼佳医生。
文礼佳医生(惹兰勿刹):先生,感谢代部长分享我们希望如何与社群联系的做法。但我觉得我们的做法相当被动防守。能否在 MUIS 内设立一个专门单位,在宗教错误信息在社群中传播开来之前,提前在互联网上搜寻这类信息?
扎吉哈先生:感谢议员的提问。首先,我认为在理想状态下,你会有无限的资源来做这件事。因为互联网非常庞大,很难找出每一个网站或每一个出现的来源。总体而言,更重要的是我们找出主题,那些能引起人们共鸣的主题;有时这些主题能引起共鸣,但同时可能是错误的,或者没有结合新加坡的语境。我们必须解释其中一些教义。
例如,几年前曾发生一些事件,有人担心是否可以祝福他人“圣诞快乐”,能不能这样祝福别人?因此,某些违背包容精神、多种族主义以及我们多宗教环境的主题,必须加以处理。
因此,我们应该主动处理其中一些主题和我们看到的新兴风险,尤其是让人们对网络伤害保持警觉。因为这些未必来自网站。生成式 AI 也不是来自网站;它来自从不同网站汇集内容的算法,因此很难监管算法。
归根结底,这与我们昨天在本院谈到的风险相同,即可能有自封的网红也在提供信息。因此,问题不在于把守信息从哪里来,而实际上也在于给我们的人民“打预防针”,让他们懂得如何识别风险,帮助家长认识孩子身上正在发生的风险。更大的图景是,我们如何帮助新加坡人,尤其是新加坡穆斯林,把许多来自外部的教义放到我们自己多种族、多宗教的语境中理解。
因此,我们需要在不同层面上努力。当然,今后我们会与社群合作。
英文原文
SPRS Hansard 原始记录 · 抓取日期:2026-09-26
12 Mr Saktiandi Supaat asked the Acting Minister-in-charge of Muslim Affairs whether MUIS will issue guidance to help Muslims who use generative AI and other online sources to provide religious answers, to verify such answers against authoritative sources and to know when and where to consult qualified asatizah.
The Acting Minister-in-charge of Muslim Affairs (Mr Zaqy Mohamad) : Mr Speaker, generative artificial intelligence (GenAI) and online platforms can complement religious learning by making information and knowledge more accessible. However, they should not be relied upon as the primary source of religious instruction or guidance, particularly on complex religious matters that require scholarly interpretation and contextual understanding. In these situations, a qualified asatizah, or other trustworthy sources of knowledge, should be consulted.
The Islamic Religious Council of Singapore (Majlis Ugama Islam Singapura or MUIS) has been reinforcing these messages through Friday sermons, public talks and the MuslimSG platform. MUIS will continue efforts to deliver practical guidance through social media and our network of community partners to strengthen digital religious literacy.
Mr Speaker : Mr Saktiandi Supaat.
Mr Saktiandi Supaat (Bishan-Toa Payoh) : Thank you, Mr Speaker. I have two supplementary questions. I thank the Acting Minister for his reply. Someone seeking religious guidance may not have the knowledge to verify an AI-generated answer, even when the references are provided. So, my first supplementary question is, could MUIS strengthen its existing channels so that members of the public can submit such answers for checking, for example, by qualified asatizah and receive timely responses, particularly on matters affecting important decisions related to religious questions on family or personal decisions.
The second supplementary question is how would MUIS ensure that the guidance reaches Muslims or Malay/Muslims who obtain most of the religious information online, as you mentioned, including those who may not regularly attend mosque programmes or religious classes? Will there be, for example, practical examples that MUIS may do, going forward, to help them recognise when an AI-generated answer needs to be checked and which ones would need to be checked by a qualified asatizah?
Mr Zaqy Mohamad : Thank you, Mr Speaker. Those are certainly valid questions and something that is of concern, not just to the Muslim community, but I think the wider community, in terms of how one consumes content today, which is available worldwide. The influences come from global influences as well.
So, I think, the issue is really about how we make our not just Jemaah, but ummah also, in terms of the wider citizen pool, better understand the need for being more aware to be able to check information, whether religious content or otherwise. I think, ultimately, that is something that we need to work on as a country.
Two things: one, to the Member's question of can we make our platforms more accessible, the answer is certainly. There is certainly room for us to improve accessibility to information, accessibility to asatizah, for example. This is one area in which MUIS is looking into, in terms of upgrading the MuslimSG platform.
GenAI is something that we look quite closely at. So, we are educating and even upgrading many of our asatizah to be more familiar with GenAI and new digital tools. Some of them have already undergone training to better understand the effects or how to use AI better, because ultimately, they do have to explain.
Secondly, I think the wider issue is how do we get the Muslim community to better engage the platforms, both physical and online platforms. Because I think the way we look at religious education has to be a holistic approach, both in the physical realm and digital realm. The asatizahs today are accessible, you can reach them via the Asatizah Recognition Scheme website. But certainly through MUIS, reach out to us if there are any concerns or any religious guidance that you seek. This is one area in which, perhaps, in the future as we upgrade MuslimSG, we could make it more interactive and more accessible for the community.
I think, ultimately, that is one area in which we are very cognisant about, in trying to reach out to the young especially, who may be influenced, especially where you have the risk of radicalisation and even what we call "ajaran sesat" or false teachings.
Within the asatizah community, I think the ability to contextualise some of the complex issues today that many get or receive or are influenced by from overseas clerics or Muftis. I think those are important aspects, in which we need to be able to help the community also contextualise many of these teachings back into the Singapore context too.
So, there are many layers that need to be worked on and that we are working on today. But I assure the Member that this is one area that we are looking at. And the online committee that I announced yesterday is also tasked to look into some of these emerging risks that come from online sources and online harms. We are certainly looking at it, both from a present state, and as well as proactively, looking at emerging risks too.
Mr Speaker : Dr Wan Rizal.
Dr Wan Rizal (Jalan Besar) : Sir, I thank the Acting Minister for sharing approaches on how we want to connect with the community. But I feel that what we do is pretty defensive. Is there some way we could create a dedicated unit, somewhere in MUIS, to comb the Internet and search for such religious misinformation in advance, before it gains traction to the community?
Mr Zaqy Mohamad : I thank the Member for his question. First, I think in an ideal state, you would have unlimited resources to do that. Because the Internet is very vast. It is very difficult to pick up every single site or every single source that comes along. Overall, what is more important is that we pick out themes, themes that resonate with people and sometimes, these are themes that resonate, but at the same time, may be false or not contextualised to Singapore's context. And we would have to explain some of these teachings.
For example, there were incidents a few years ago, where there were concerns of whether you should be able to wish others "Merry Christmas", can you wish someone that? So, certain themes which goes against the spirit of inclusivity, multi-racialism, our multi-religious context, have to be addressed.
So, we should proactively work on some of these themes, emerging risks that we see, and especially in terms of being able to keep people aware of online harm. Because they may not necessarily come from sites. GenAI does not come from sites as well; it comes from algorithms that are pulled together from different sites, so it is very difficult to police algorithms.
Ultimately, it is also about the same risk as we spoke about yesterday in this House, where you could have a self-styled influencer giving information too. So, it is not about guarding where the information comes from, but it is really also about inoculating our people, in terms of how to recognise the risk, helping parents to recognise the risk with what is going on with their kids. The broader picture is how do we help Singaporeans, especially Singaporean Muslims, contextualise many of these teachings, from outside, into our own multiracial, multi-religious context too.
So, there are different layers at which we need to work on. And certainly, we will work with the community as we move forward.