SEIZING AGENTIC AI OPPORTUNITY IN M’SIA

 

Leading companies are moving beyond experimentation as a third of “AI future-built” firms have deployed agentic solutions and are demonstrating measurable value.

MALAYSIA stands at a critical inflection point in the global artificial intelligence (AI) race.

After the surge of generative AI, a new wave is emerging in the form of agentic AI.

Agentic AI are AI execution models involving autonomous agents that coordinate across workflows, tools and systems with minimal human input.

While it stops short of true autonomous decision-making, agentic AI’s ability to make actionable decisions within predefined parameters is a game changer.

Malaysia has a solid foundation to build on.

According to Boston Consultant Group’s AI Maturity Matrix, which benchmarks 73 economies globally on AI exposure and AI readiness, Malaysia is classified as a “steady contender”.

It places the nation just one tier behind AI pioneers such as the United States, the United Kingdom and China.

This position reflects Malaysia’s significant exposure to AI, particularly in large sectors like retail and wholesale, telecommunication and financial services.

At the same time, it indicates a solid level of AI readiness, supported by forward-looking ambitions, policies and regulatory frameworks on AI.

An evolving AI landscape

AI is rapidly becoming a critical national infrastructure that empowers wider opportunities.

As a result, geopolitical shifts, compute access and sovereign capability increasingly determine economic outcomes and geopolitical influence.

The US and China lead the global AI race.

Tech companies from these two superpowers created 59% and 26%, respectively, of top-performing large language models (LLM).

This presents a conundrum for competing nations.

Relying solely on external technology providers poses challenges for corporate leaders and governments, especially since local regulations, data requirements and model availability are subject to shifting policies.

Against this backdrop, a small group of “GenAI middle powers” is emerging across Europe, Asia and the Middle East.

Each has distinct strengths that might allow it to compete as a regional or global technology supplier.

This race now expands beyond software to encompass hardware, infrastructure and technology adoption.

Malaysia must actively build its domestic AI capabilities to avoid high technology sovereignty risks as it looks to the future of agentic AI.

Execution speed and scale will dictate whether Malaysia leads in Asean or falls behind.

Encouragingly, the Digital Ministry, through the establishment of the National AI Office (NAIO), is driving a coordinated national AI agenda – spanning governance frameworks, cross-sector adoption and ecosystem development.

These efforts lay the critical foundations for more advanced applications, including the next wave of agentic AI.

Productivity multiplier

Globally, the shift is already underway and early signs indicate that the rise of agentic AI will be rapid.

BCG’s Build for the Future 2025 study shows that agentic AI’s share of AI-driven value is expected to nearly double from 17% in 2025 to 29% by 2028.

Leading companies are moving beyond experimentation – one-third of “AI future-built” firms have deployed agentic solutions and are demonstrating measurable value.

Early adopters are already unlocking tangible benefits. BCG’s study shows that while companies are exploring agentic AI across operations, support functions and innovation, customer experience is emerging as the top priority.

Leading use cases include deploying intelligent agents to autonomously handle Level 1 and Level 2 customer support, as well as optimising digital marketing campaigns – continuously adjusting bids to maximise returns, reallocating spend to high-performing channels and testing creatives in real time.

AI undoubtedly represents a powerful productivity multiplier for Malaysia.

It can strengthen key economic sectors such as manufacturing, financial services and many other industries. For SMEs, agentic AI can lower the cost of sophistication, providing access to capabilities once reserved for large enterprises.

Beyond the private sector, agentic AI can modernise public services and improve policy-making decisions and delivery in healthcare, education and justice.

It can help bridge urban-rural divides by expanding access to digital services and decision support.

In a nation balancing growth ambitions with demographic and fiscal constraints, agentic AI is not merely a technology upgrade – it is a lever for sustainable and inclusive growth.

Four strategic priorities

To compete effectively in this next phase of AI, Malaysia must act with clarity and intent across four priorities.

> Build sovereign AI capabilities. Malaysia could strategically build sovereign AI capabilities in areas where it has natural strengths and where risk mitigation matters most.

This includes expanding reliable access to compute, leveraging its growing data centre ecosystem.

A pragmatic and technology-neutral approach that combines global technology partnerships with targeted domestic capability-building will be more effective than pursuing full-stack independence.

Technology partnerships could focus on leveraging leading AI innovations from both Western and Eastern ecosystems in a neutral manner.

Open-source AI models offer a practical pathway to reduce dependency risks, accelerate adoption and support local customisation.

At the same time, efforts could focus on enabling responsible use of high-quality local datasets.

> Invest aggressively in talent. Malaysia must pair global talent attraction with sustained local capability development to build the AI workforce needed to compete at scale.

It could aggressively attract top global AI talent through competitive incentives, strong research ecosystems and vibrant innovation hubs, while simultaneously building a deep domestic pipeline of AI talent.

This requires strengthening STEM education, expanding university–industry collaboration, embedding AI in technical and vocational training and accelerating workforce upskilling across sectors.

> Scale national platforms. Malaysia must move from fragmented pilots to scaled national platforms, anchored on high-impact use cases – such as a unified government interface linked to MyDigitalID.

This platform provides a common foundation to embed AI agents that deliver personalised public services.

Scaling up such platforms will catalyse greater private-sector participation and ensure sustainable adoption of agentic AI.

In addition, Malaysia could strengthen exchange platforms that bring together the government, industry and academia to accelerate collaboration, capability-building and use case development.

Associations such as AI Malaysia (AIM), Malaysian Autonomous Intelligence & Robotics Association (MyAIRA), along with other industry associations, can play a critical role in sharing best practices, mobilising talent and aligning stakeholders to drive ecosystem-wide adoption of agentic AI.

> Implement pro-innovation regulation. Malaysia needs regulations that protect users but also preserve competition.

Policymakers could favour a flexible model over rigid frameworks, particularly in a fast-evolving technological landscape.

Malaysia could pursue a balanced approach – combining principle-based guidelines, regulatory sandboxes and sector-specific standards that can evolve alongside the technology.

Priming Malaysia for growth is critical, but it is essential that this is done through a forward-looking and ethical approach.

Malaysia has the opportunity to differentiate itself by championing ethical, inclusive AI.

This is a core foundation of effective AI adoption, and should align with national values, ensuring that trust and confidence underpin the next wave of innovation in agentic AI.

Defining the future

The stakes are clear. AI investment compounds rapidly. Early movers attract capital, talent and vibrant ecosystems.

The choice is not whether AI will reshape the Malaysian economy.

The choice is whether Malaysia will shape that transformation with speed, clarity and ambition while remaining anchored to core Malaysian values.

CF Ong is managing director and senior partner in Boston Consulting Group.

How A.I. Chatbots Like ChatGPT and DeepSeek Reason

 

An illustrative image showing the artificial intelligence chatbot ChatGPT on a smartphone in San Francisco. In September of last year, Open AI released a new “reasoning” version of its ChatGPT chatbot that was designed to spend time “thinking” through complex problems before settling on an answer. Now other companies like Google, Anthropic and China’s DeepSeek offer similar technologies. — KELSEY MCCLELLAN/NYT

In September, OpenAI unveiled a new version of ChatGPT designed to reason through tasks involving math, science and computer programming. Unlike previous versions of the chatbot, this new technology could spend time “thinking” through complex problems before settling on an answer.

Soon, the company said its new reasoning technology had outperformed the industry’s leading systems on a series of tests that track the progress of artificial intelligence.

Now other companies, like Google, Anthropic and China’s DeepSeek, offer similar technologies.

But can AI actually reason like a human? What does it mean for a computer to think? Are these systems really approaching true intelligence?

Here is a guide.

What does it mean when an AI system reasons?

Reasoning just means that the chatbot spends some additional time working on a problem.

“Reasoning is when the system does extra work after the question is asked,” said Dan Klein, a professor of computer science at the University of California, Berkeley, and chief technology officer of Scaled Cognition, an AI startup.

It may break a problem into individual steps or try to solve it through trial and error.

The original ChatGPT answered questions immediately. The new reasoning systems can work through a problem for several seconds – or even minutes – before answering.

Can you be more specific?

In some cases, a reasoning system will refine its approach to a question, repeatedly trying to improve the method it has chosen. Other times, it may try several different ways of approaching a problem before settling on one of them. Or it may go back and check some work it did a few seconds before, just to see if it was correct.

Basically, the system tries whatever it can to answer your question.

This is kind of like a grade school student who is struggling to find a way to solve a math problem and scribbles several different options on a sheet of paper.

What sort of questions require an AI system to reason?

It can potentially reason about anything. But reasoning is most effective when you ask questions involving math, science and computer programming.

How is a reasoning chatbot different from earlier chatbots?

You could ask earlier chatbots to show you how they had reached a particular answer or to check their own work. Because the original ChatGPT had learned from text on the internet, where people showed how they had gotten to an answer or checked their own work, it could do this kind of self-reflection, too.

But a reasoning system goes further. It can do these kinds of things without being asked. And it can do them in more extensive and complex ways.

Companies call it a reasoning system because it feels as if it operates more like a person thinking through a hard problem.

Why is AI reasoning important now?

Companies like OpenAI believe this is the best way to improve their chatbots.

For years, these companies relied on a simple concept: The more internet data they pumped into their chatbots, the better those systems performed.

But in 2024, they used up almost all of the text on the internet.

That meant they needed a new way of improving their chatbots. So they started building reasoning systems.

How do you build a reasoning system?

Last year, companies like OpenAI began to lean heavily on a technique called reinforcement learning.

Through this process – which can extend over months – an AI system can learn behavior through extensive trial and error. By working through thousands of math problems, for instance, it can learn which methods lead to the right answer and which do not.

Researchers have designed complex feedback mechanisms that show the system when it has done something right and when it has done something wrong.

“It is a little like training a dog,” said Jerry Tworek, an OpenAI researcher. “If the system does well, you give it a cookie. If it doesn’t do well, you say, ‘Bad dog’.”

(The New York Times sued OpenAI and its partner, Microsoft, in December for copyright infringement of news content related to AI systems.)

Does reinforcement learning work?

It works pretty well in certain areas, like math, science and computer programming. These are areas where companies can clearly define the good behavior and the bad. Math problems have definitive answers.

Reinforcement learning doesn’t work as well in areas like creative writing, philosophy and ethics, where the distinction between good and bad is harder to pin down. Researchers say this process can generally improve an AI system’s performance, even when it answers questions outside math and science.

“It gradually learns what patterns of reasoning lead it in the right direction and which don’t,” said Jared Kaplan, chief science officer at Anthropic.

Are reinforcement learning and reasoning systems the same thing?

No. Reinforcement learning is the method that companies use to build reasoning systems. It is the training stage that ultimately allows chatbots to reason.

Do these reasoning systems still make mistakes?

Absolutely. Everything a chatbot does is based on probabilities. It chooses a path that is most like the data it learned from – whether that data came from the internet or was generated through reinforcement learning. Sometimes it chooses an option that is wrong or does not make sense.

Is this a path to a machine that matches human intelligence?

AI experts are split on this question. These methods are still relatively new, and researchers are still trying to understand their limits. In the AI field, new methods often progress very quickly at first, before slowing down. – ©2025 The New York Times Company

This article originally appeared in The New York Times.

Related:

STaR: The AI That Teaches Itself to Reason

Medium · Sahin Ahmed, Data Scientist

10+ likes · 7 months ago

STaR is a pioneering step toward AI that not only performs better but also learns, reasons, and explains itself in ways that are deeply aligned with human .

AI seen as key to high-quality growth

Chinese expert urges Western countries to view rise of DeepSeek fairly, as firm faces scrutiny

 

DeepSeek

A Chinese expert on Friday called on Western countries to view the rise of Chinese technology firms fairly, as Chinese artificial intelligence (AI) startup DeepSeek faces growing scrutiny from some countries following its rise in the international market. 

Italy's data protection authority announced Thursday it had blocked DeepSeek's AI services, citing insufficient transparency about how the company handles personal information. The move, reported by Reuters, follows Ireland's Data Protection Commission's request for details from DeepSeek on how it processes Irish user data.

France's privacy watchdog, the CNIL, said on Thursday that it plans to investigate DeepSeek's AI system operations and potential privacy risks. "The CNIL's AI department is currently analyzing this tool," a CNIL spokesperson said on Thursday, according to Reuters.

Technology should be shared and used for the benefit of all humanity, rather than being monopolized, Liu Dingding, a veteran tech industry analyst, told the Global Times on Friday. 

"Western countries should view the success of DeepSeek's R1 AI model fairly, allowing more people to enjoy the conveniences brought by technological advancement, instead of imposing restrictions," Liu Dingding said.

Meanwhile, DeepSeek is gaining traction in the international market, including adoption by some US tech giants. US-based chipmaker Nvidia announced on Thursday that the Chinese startup's model is now available on its NIM platform for enterprises to deploy, highlighting its growing acceptance in the US market. 

Microsoft revealed on Wednesday that it has added DeepSeek's R1 AI model to its Azure cloud platform and GitHub developer tools.

Apple CEO Tim Cook said that DeepSeek's AI models are "innovation that drives efficiency," when asked about the Chinese AI startup during an earnings call on Thursday. "In general, I think innovation that drives efficiency is a good thing. And, you know, that's what you see in that model," said Cook, according to Yahoo Finance.

The tech community, including those directly involved in AI development and academia, acknowledges DeepSeek's innovations as substantial milestones, Liu Wei, director of the human-machine interaction and cognitive engineering laboratory with the Beijing University of Posts and Telecommunications, told the Global Times on Friday.

However, for some Western countries, China's sudden emergence of groundbreaking achievements in the AI field, as exemplified by DeepSeek's rise, may have caught them off guard, and they need to find ways to respond to China's technological progress, said Liu Dingding.

The so-called restrictive measures, often justified by so-called concerns over personal information protection, to some extent reflect these countries' attitudes toward China's rise in technology, Wang Peng, an associate research fellow at the Beijing Academy of Social Sciences, told the Global Times on Friday.

However, as an open-source platform, DeepSeek's development will not be significantly hindered by these measures, said Wang,

In the US, two congressmen raised so-called national security concerns by urging the Trump administration to restrict exports of Nvidia's advanced AI chips to China, singling out DeepSeek as a key recipient of these chips. 

Their letter urged US National Security Advisor Michael Waltz to "consider the potential national security benefits of placing an export control on Nvidia's H20 and chips of similar sophistication," Reuters reported on Friday.

It is unlikely that these restrictive measures will succeed in containing China's technological development, as the rise of DeepSeek's AI model serves as evidence, according to Liu Dingding.

In Asia, Japan's Chief Cabinet Secretary Yoshimasa Hayashi took a measured approach during a Tuesday press conference, noting that no specific regulations targeting DeepSeek had been enacted but affirming that Japan would closely track global AI trends, Nikkei reported on Thursday.

Source link

How to Download and Use DeepSeek on PC and Mobile for Free: Step-by-Step Guide

 

How Malaysians can learn to love AI

How Malaysians can learn to love AI

DIGITAL initiatives seem to be coming fast and furious from the government.

We had the launch of the Central Database Hub, or Padu, on Jan 2 and just this week, on Tuesday, the AI for the Rakyat portal, was unveiled.

AI untuk Rakyat

AI For Rakyat

, https://ai.gov.my/#/home

The introduction at its website, ai.gov.my, is interesting: “AI untuk Rakyat” is a self-learning online programme designed to raise public awareness about artificial intelligence (AI).

“It aims to demystify AI for people from all walks of life – a student, a stay-at-home parent, a professional in any field, a senior citizen – basically, anyone interested in getting acquainted with it and wanting to build a ‘Digital First Mindset’.”

If all goes well, the portal could indeed help to address the country’s growing digital divide as those with access to the digital world pull away from those who don’t have the means to benefit from it.

Padu is entirely designed and maintained by the government to ease fears of our data being misused in the hands of a third party.

Similarly, AI for the rakyat is largely government-driven, with the Economy Ministry working with just one private company, Intel.

And we are glad to note that apart from Bahasa Malaysia, the portal is also inclusively available in English, Chinese and Tamil (the last is rare indeed).

The government is hoping to have one million Malaysians become familiar with AI skills in three years.

That’s all well and good, but we must ensure that this portal does not go the way of previous tech initiatives, like the Multimedia Super Corridor, which did not reach its full potential.

All these digital initiatives need real support in the form of nurturing ecosystems and infrastructure, serious investment, and an open acceptance of all sorts of different talents.

Most importantly, “knowing” AI does not lessen the urgent need for solid cybersecurity laws and a moral code of ethics to govern the use and development of all digital initiatives, including AI.

As Economy Minister Rafizi Ramli stressed at the portal’s launch, there must be optimal end-to-end conditions and wholesome policies to accompany the growth of AI.

Rafizi argued that without trustworthy data and secure channels, the products and services driven by AI technology could not be implemented properly.

“To be the regional and global AI hub, we must be able to create new uses for AI. This can only be done if we have ready and accessible data,” he said, pointing out that Padu is a step towards this.

AI untuk Rakyat

AI For Rakyat

, https://ai.gov.my/#/home

https://ai.gov.my/#/home

About AI untuk Rakyat

AI untuk Rakyat is a self-learning online program designed to raise public awareness about Artificial Intelligence (AI). It aims to demystify AI for people from all walks of life – a student, a stay-at-home parent, a professional in any field, a senior citizen - basically, anyone interested in getting acquainted with it and wanting to build a ‘Digital First Mindset' 

The program is divided into two sections: AI Aware and AI Appreciate. 

Both sections can be completed in about four hours. Each section is based on AI-related concepts explained through engaging activities and quizzes. The users can take the quizzes as many times as they want. After they complete these quizzes, they become bearers of AI Aware and AI Appreciate badges that they can share on their social media accounts.

What is AI untuk Rakyat?

Who is this program for? Who should/ can attend/ participate in it?

What is the duration of this program?

What is covered in the program?

What are the prerequisites for this program?

What will I get once I complete the program?

Can I call myself an AI expert after attending this program?

I have finished the program. What is the next step?

I am unable to select my birth year on the registration page. What should I do?

I did not receive the OTP. What do I do?

My query is not resolved yet - whom should I reach out to?