Singapore introduces three new AI governance initiatives

On Monday, the Singapore government introduced three new artificial intelligence (AI) governance initiatives to enhance safety for use both locally and globally. 

The products were the Global AI Assurance Pilot for best practices around technical testing of generative AI (GenAI) applications; a joint testing report with Japan; and the publication of the Singapore AI Safety Red Teaming Challenge evaluation report.  

The announcement was made by Singapore’s Minister for Digital Development and Information, Josephine Teo, at the AI Action Summit (AIAS) held in Paris, France. 

The AIAS summit was organised on February 10 and 11. Various global leaders, like French President Emmanuel Macron, Indian Prime Minister Narendra Modi, Chinese Vice Premier Zhang Guoqing and US Vice-President, JD Vance, were among the dignitaries attending the event.

Speaking at an AIAS panel on Monday, Minister Teo said the three new initiatives reflect Singapore’s commitment to rallying industry and international partners towards concrete actions that advance AI safety. 

The Minister noted that there was increasing recognition that trust is the foundation of widespread AI adoption. “Without it, there’ll be reduced confidence that AI can be relied upon to transform businesses and improve lives,” she said. 

She noted that to fully harness AI for the Public Good, Singapore believes that it must first fully understand what people distrust about AI (such as bias in systems and loss of control); and find fair and rigorous ways to test AI systems so they are safer and more responsible. 

To subscribe to the GovInsider bulletin click here.

In her speaking engagements, she acknowledged that within the Singapore government, “some things are well known, and some things are not well understood”.  

“In particular, when it comes to AI and what kind of use cases and applications that will have the most receptivity to the market, that is something that we think the private sector knows a lot more,” Minister Teo said.  

So, with that as a starting point, we wanted very much to get the private sectors’ viewpoints, and to understand what kinds of signal they respond to best, she added.  

“And on that basis, I would say that the private sector partnership has turned out to be very encouraging,” Minister Teo said. 

How Singapore can help 

When asked how Singapore could help other countries through partnerships, the Minister said, “We have learning lessons, some of our own experiences in developing public good use cases, as well as risk management measures, [that] we are happy to share with colleagues around the world.”  

The Minister said she was “most encouraged” by the Network of AI Safety Institutes as they were not “just a platform where the AI safety institutes are exchanging notes. I think they are taking joint action”. 

During a panel discussion on Sunday, the Minister also noted that the government is looking at the use of generative AI (GenAI) to identify scams and do a better job at law enforcement. 

Minister Teo noted that the Global AI Assurance Pilot, launched by the AI Verify Foundation and the Infocomm Media Development Authority (IMDA), has been established as a testbed for the best global practices around technical testing of GenAI applications.  

According to Singapore’s Ministry for Digital Development and Information (MDDI), the pilot will convene leading AI assurance and testing vendors with firms which are deploying real-life GenAI applications.  

It will shape future AI assurance standards and future assurance services, grow the local and international third-party AI assurance markets, and provide practical input to AI governance frameworks, MDDI said. 

Guardrails for LLMs 

The Minister noted the release of a joint testing report in collaboration with Japan under the AI Safety Institute (AISI) Network, which aimed to make large language models (LLMs) safer in different linguistic environments by assessing whether guardrails hold up in non-English. 

As co-lead of the Testing and Evaluation Track under the AISI network, Singapore has brought together global linguistic and technical experts from the AISI network to conduct tests across 10 languages (Cantonese, English, Farsi, French, Japanese, Kiswahili, Korean, Malay, Mandarin Chinese, Telugu) and five harm categories (violent crime, non-violent crime, IP, privacy, jailbreaking) to build up evaluation capabilities and methodological standards.  

The joint testing exercise expanded on global efforts to make models safer in different linguistic environments, given current English-centric training and testing which potentially leaves gaps in non-English safeguards.  

Red teaming challenge 

Minister Teo added that the Singapore AI Safety Red Teaming Challenge Evaluation Report 2025 was published to get a better understanding of how LLMs perform with regard to different languages and cultures in the Asia Pacific region, and if the safeguards hold up in these contexts.  

The report sets out a consistent methodology so that “we can test across diverse languages and cultures, as no one party can accomplish that alone”, MDDI said.  

The report is based on findings from the AI Safety Red Teaming Challenge, organised by IMDA and Humane Intelligence, a non-profit testing organisation, in November 2024.  

More than 50 participants from nine countries across the Asia Pacific came together and red-teamed four LLMs (Aya, Claude, Llama, SEA-LION) for cultural bias stereotypes in non-English languages, compared to English.  

The challenge aimed to advance the sciences in AI testing, a nascent space globally. The data collected will be used to develop benchmarks and automate testing for regional safety concerns. 

Minister Teo noted that AI is reshaping our world. “Singapore sees an opportunity to contribute to AI safety and governance. We are committed to strong global partnerships — with industry, academia, and community — so that we realise AI for the Public Good for Singapore and the world.” 

MDDI said in a statement that Minister Teo met with policymakers, industry leaders, and academics at the sidelines of AIAS, “exchanging insights on AI safety, regulatory frameworks, and emerging AI trends”.