Advertisement

Many businesses approach artificial intelligence (AI) as they would conventional computing — by purchasing ready-made solutions from tech giants such as OpenAI, Microsoft or Google and plugging them into their operations. But this mindset is flawed, says Goh Peng Ooi, president of AI Malaysia (AIM), who warns that such assumptions often lead to costly missteps.

“Many people don’t realise that AI is actually not part of conventional computing … If you think so, you’re going to waste a lot of money and time, you are going to lose a lot of money,” says Goh, a tech veteran and founder of fintech firm Silverlake Axis. He is also widely regarded as Malaysia’s first tech billionaire.

Launched on May 30, the Asean Business Advisory Council (Asean-BAC)-backed AIM is to act as a platform for the private sector to discuss and share resources for an effective nationwide adoption of AI solutions.

The country is already in a key position for this collective approach. According to the Boston Consulting Group’s (BCG) AI Maturity Metrics Report published last November, Malaysia is among the top 50% of countries in terms of AI readiness, and is among the three economies in Southeast Asia (along with Singapore and Indonesia) that are the front runners in AI.

For Goh, addressing this misconception of treating AI like conventional computing, email or cloud storage is just as important. AI is not just a tool to implement, he says. Rather, it is an entire system that a company can develop itself.

Drawing a comparison with smartphones, Goh notes that while the technology behind them is expensive and complex, it is not economically viable for a Malaysian company to build its own. Instead, it makes more sense to act as a vendor, selling devices from established brands such as Samsung or Apple.

Similarly, in the AI space, platforms such as ChatGPT and Gemini are equivalent to these global tech giants. Goh wants to raise awareness that unlike with smartphones, local companies can adopt AI affordably by not relying on these big brands. He explains that AI can be developed using two main approaches: backpropagation and reinforcement learning.

To simplify, backpropagation has the AI model look back at its process when a mistake is made, calculate this mistake and adjust its answers accordingly. OpenAI’s ChatGPT is built on this method.

Reinforcement learning, which China’s Deepseek uses, has an AI agent that adjusts its decisions and output based on what it learns from its environment, based on rewards and penalties.

Goh points out that backpropagation requires huge capital to develop and puts data in the control of companies such as Microsoft and Google. By contrast, reinforcement learning is cheaper to develop and using it can help Malaysia become its own vendor for AI.

By dispelling this myth, Goh hopes that not only will AI adoption in the private sector grow, but it will also empower Malaysian companies to take control of their data. He argues that this will not only secure data sovereignty but also prevent another colonisation, where big corporations control Malaysian data.

Connecting the private sector with stakeholders

AIM was established when Asean-BAC was looking at key projects to champion during Malaysia’s tenure as chair of Asean. Tan Sri Nazir Razak, chairman of Asean-BAC Malaysia, says AI is an area that cannot be ignored; yet, there was no proper private sector body that companies or the government could engage with in this space.

Although the National AI Office (NAIO) does reach out to the private sector to explain the benefits of AI, there was still a need for a platform where the private sector could connect with stakeholders to establish partnerships or funding processes.

“While government bodies like NAIO are important in setting up policy and infrastructure, the true engine of innovation and implementation lies with the vibrant private sector,” says Nazir.

Beyond just knowledge sharing, AIM is to serve as the main platform to foster collaborations with AI solution providers and guide companies to where they can find the required funding for their AI efforts.

“We have an association of banks, we have many chambers of commerce. It’s a very established thing that when you actually pull together the private sector, it brings collective intelligence together to advise the government,” says Nazir.

Ching-Fong Ong, AIM secretary as well as managing director and senior partner at Boston Consulting Group, adds that AIM is considering offering different types of membership. 

Companies from significantly affected sectors such as agriculture and healthcare may have their own tiers, with dedicated curation and screening processes to help members connect with relevant partners.

A basic membership for all companies may also be offered in the future. While fees have yet to be finalised, there will be a joining fee, as AIM is a non-profit entity. To encourage membership, AIM will organise free events and talks to raise awareness and understanding of AI.

AIM has extended early membership offers to council member companies such as CIMB Bank, AirAsia, Gentari and Carsome. At its launch, AIM counted 15 members — at least one from each state — though their names have not been disclosed.

At present, many of AIM’s systems remain under development, including its membership fees, overall strategy and organisational structure.

Leading by example

As a key initiative under Malaysia’s Asean chairmanship, AIM also serves a broader purpose by laying the foundation for a region-wide platform to accelerate AI adoption.

The long-term vision is to establish an Asean AI Network — a collaborative platform that connects private-sector AI associations across Asean member states to share knowledge, expertise and resources in the AI space.

“This is akin to the same work we have done on the carbon side, the Asean Common Carbon Framework, where Asean-BAC Malaysia set up the Malaysia Carbon Market Association at COP-29, signing an agreement with its counterparts in Indonesia, Thailand and Singapore,” says Nazir.

“We couldn’t get all countries on board, as not all [Asean] countries have one, but those four are now working together towards this Asean common market. So, AIM is going to be similar.”

He adds that Asean member states are supportive of AIM, noting that a central private-sector platform for AI adoption has long been needed.

Malaysia is not the first country to establish an association such as AIM — Singapore and Indonesia have their own in AI Singapore and Indonesia AI. What sets AIM apart, according to Nazir, is its emphasis on broader, consolidated regional collaboration.

To accommodate the regional, cultural and policy differences among Asean member states, the Asean AI Network will function more as a collective of these national associations rather than a single, centralised entity.

“AIM is an association. For now, the Asean AI Network will be more of a group that gets together during Asean summits and the like … Asean is a coalition that is a bit looser [unlike the European Union. So, the Asean AI network will be looser, too,” says Ong.

What this means is that the Asean AI Network will not have the power to influence policy among Asean nations, but can instead act as a platform for knowledge sharing.

“So, in a way, it’s also a platform for those more deve­loped Asean countries to help the less developed ones. It’s not that we are coming together to tell you what you have to do, [such as opening up your markets,” says Nazir.

Source: https://theedgemalaysia.com/node/758022