Flag errors in Malaysia a 'warning' to take AI governance more seriously: Experts

KUALA LUMPUR: The recent spate of botched depictions of Malaysia’s flag shows the country needs clearer regulations on the use of artificial intelligence (AI), particularly when it involves subjects of national importance, said experts.  

In two of at least four incidents this month involving inaccurate depiction of the Jalur Gemilang – which means Stripes of Glory – AI was reportedly used to generate the images.

In the first of the blunders, the front page of newspaper Sin Chew Daily’s Apr 15 print edition showed Malaysia’s flag without the crescent moon, which symbolises Islam, the country’s official religion.

The other incident involved the Education Ministry, whose examination results analysis report distributed on Apr 24 showed the Jalur Gemilang with two stars instead of one, as well as eight alternating red and white stripes instead of 14. The stripes symbolise the equal status of Malaysia’s 13 states and federal territories, while the star symbolises their unity.

Subscribe to CNA’s Morning Brief

An automated curation of our top stories to start your day.

This service is not intended for persons residing in the E.U. By clicking subscribe, I agree to receive news updates and promotional material from Mediacorp and Mediacorp’s partners.

Loading

The incidents show Malaysia’s public institutions and private sector have readily adopted AI, but often overlook its flaws and potential harms, analysts said.

The lapses serve as a “warning” to take AI governance more seriously, especially in sensitive areas such as national symbols, said media analyst Nuurrianti Jalli of Singapore’s ISEAS-Yusof Ishak Institute.

NOT ONLY THE JALUR GEMILANG

But it is not only AI images involving national emblems that have stoked public disquiet and led to calls for better regulation.

On Apr 19, for instance, an image allegedly generated by AI of a durian farm surrounded by forest triggered speculation over the issue of illegal farmers in Malaysia’s Pahang state. 

According to the South China Morning Post, the image’s caption suggested that it was taken at a “hidden spot” and thus fuelled public belief that there were other illegal farms hidden in the area’s dense forests. 

The Malaysian Communications and Multimedia Commission questioned the man who uploaded the image.

Other cases involved fake images showing the supposed “arrest” of the country’s monarch Sultan Ibrahim, as well as sexual content created with deepfake technology. 

HUMAN OVERSIGHT NEEDED

Malaysia has been actively promoting AI adoption under initiatives such as “AI for the People”, launched in 2024 by Prime Minister Anwar Ibrahim as part of a broader push to close the digital literacy gap among Malaysians.

In the wake of the national flag incidents, Anwar said that while he supports the use of AI across various sectors such as media, education and business, the technology “must not replace human editorial judgment and quality control”. 

His press secretary said investigations would be carried out. Both Sin Chew Daily and the Education Ministry have apologised, and the Chinese daily also suspended two of its editors and vowed to strengthen its editorial processes involving AI. 

AI systems require human oversight and control, said Elina Noor, a senior fellow at the Asia Programme at Carnegie Endowment for International Peace. “Just because AI is used doesn’t mean that human oversight should be absolved.”

Users must understand “invisible risks” such as bias embedded in the design of AI systems that can cause “visible risks” in the form of misinformation, which could result in societal harm, said Elina.

“There is a tendency in the region to treat AI solutions as neutral, efficient and convenient because they are viewed as technical instruments,” she said.

But there is growing awareness that AI models can contain cultural, gender and other biases. 

A study involving four large language models in late 2024 by Singapore’s Infocomm Media Development Authority, in partnership with AI auditing firm Humane Intelligence, found that racial stereotypes were reinforced.

There were also generalisations about people from different parts of a country, such as that people from the north of Malaysia held more traditional beliefs on marriage and festivals, while those in the south were more liberal.

WHAT RULES TO GOVERN AI USE?

Organisations should spell out policies on the proper use of AI and ensure compliance, said Elina.

Regulations should mandate that declarations be made when AI has been used in artwork or videos on official platforms, for instance, said communications lecturer Benjamin Loh from Monash University Malaysia.

The authorities should provide specific details about the laws regarding AI-generated content, addressing whether AI works can be considered copyright, and whether AI systems may use copyrighted works as training data without permission under fair-use provisions, Loh said. 

There should also be rules protecting those in the creative industry, he said.

Policies must, however, strike a balance so that they do not stifle innovation, experts said.

“We don’t want to shut down innovation or make it too difficult for companies to use AI,” said Nuurrianti, a visiting fellow at the ISEAS’s Media, Technology and Society Programme.

“It’s not about choosing between growth and regulation as we have to do both. If we get the balance right, Malaysia could become a leader in Southeast Asia, not just in building AI technologies, but in setting an example for responsible and ethical AI development,” she added. 

Malaysia has secured billions of dollars in investment from global tech firms including Google and Amazon, which have all announced data centre and AI projects in the country. 

At the regional level, the Association of Southeast Asian Nations (ASEAN), which Malaysia chairs this year, has taken some steps towards aligning regulation of AI.

The bloc issued guidelines on AI governance and ethics in February last year to serve as a “practical guide for organisations in the region that wish to design, develop, and deploy traditional AI technologies in commercial and non-military or dual-use applications”.

However, Nurriianti said some of these regional mechanisms are still in their “early days” and will depend on the risks, capacity and context of each country. 

In the meantime, Malaysia should forge ahead with clear national policies for responsible use of AI, especially in public-facing applications, she said.

In December last year, Anwar launched Malaysia’s National Artificial Intelligence Office aimed at shaping policies and addressing regulatory issues, as the country looks to establish itself as a regional hub for AI development. CNA has reached out to the office for comments on the recent flag incidents. 

“We are reaching a point where AI is not just a tool but it’s also shaping how people see their country and themselves,” said Nurriianti.

“As AI gets more powerful, the risks get bigger too, especially in areas tied to identity, politics and social trust,” she said.