This article first appeared in Digital Edge, The Edge Malaysia Weekly on March 24, 2025 - March 30, 2025

It is no secret that artificial intelligence (AI) is being touted as a tech superstar with the potential to redefine the way many industries currently operate. In Malaysia’s Budget 2025, the government announced a significant allocation of RM50 million to bolster the teaching of AI in research universities.

The investment underscores the country’s commitment to advancing its technological capabilities and also in preparing its workforce for the challenges of an increasingly AI-driven world, says Febrianto Siboro, managing director for Malaysia, Indonesia and Vietnam at SAS, a global provider of AI and analytics software.

“As Budget 2025 accelerates Malaysia’s transformation in AI, local businesses must balance promoting innovation with upholding moral principles and guaranteeing transparency as AI legislation develops internationally. Companies can adopt ethical standards and conduct frequent audits that check AI for bias, fairness and privacy concerns in order to comply with changing rules,” says Siboro.

With the rapidly evolving global AI landscape, Siboro shares insights from SAS’s 2025 AI predictions, which shed light across several fields, offering perspectives on how local businesses can navigate the landscape and stay competitive.

The environmental impact of AI certainly cannot be ignored, stresses Siboro. As such, companies must foster innovation while aligning with environmental, social and governance (ESG) goals.

The rapid evolution of AI, fuelled by energy-intensive cloud computing, has heightened the urgency for sustainable practices. Large-scale AI models require immense computational resources, driving energy consumption and increasing emissions. Methods such as model compression and quantisation are emerging as practical solutions, enabling companies to reduce energy use by optimising the size and complexity of AI models.

The struggle also lies, however, in the fact that many Malaysian businesses often consider environmental responsibility to lie primarily on the cloud vendor — but the truth is that this is a shared accountability.

“Both cloud providers and consumers will probably actively participate in sustainable practices as AI workloads increase. While customers are urged to embrace effective AI models and appropriately manage compute demands, providers are using renewable energy and optimising data centres to reduce emissions,” says Siboro.

Early consideration of AI ethics is essential for avoiding abuse and directing the deployment of AI in an ethical manner. As such, he says, Malaysian businesses should “begin developing ethical AI by making sure that the training data is diverse and of high quality, eliminating biases that could have negative effects”.

For continuous supervision, with accountability for mistakes and frequent audits to avoid unintentional bias, strong governance with defined roles — such as a chief AI ethicist — is crucial. To modify training data, data sources and algorithm behaviour, ethical frameworks must permit human intervention.

“Human input helps AI systems become more accurate and equitable by addressing potential biases. This ensures that AI responsibly meets the needs of a wide range of users,” says Siboro.

Established ethical standards and human oversight are essential for identifying and addressing errors in AI decision-making, while advanced cybersecurity measures can help businesses prevent their AI systems from being exploited by cybercriminals.

“Local businesses should make sure that sensitive data and models kept in virtual repositories are safe from tampering or unwanted access. To identify risks early, it’s also critical to regularly check for odd inputs or behaviours that can point to manipulation or hostile actions.

“Companies should use AI-driven security solutions to continuously identify and eliminate any potential flaws or exploits that target AI systems. These measures will lessen the increasing dangers of cybercriminals abusing AI,” he says.

Collaboration across industries and continuous AI training are critical to adapting to evolving threats. In addition, initiatives such as Malaysia’s National Fraud Portal further enhance the national’s resilience, enabling quicker responses to financial fraud and protecting businesses from cybercriminal exploitation.

As generative AI (Gen AI) continues to attract attention, the test for businesses lies in moving beyond the hype to achieving measurable outcomes, says Siboro.

According to the IDC Data and AI Pulse: Asia Pacific 2024 report, which SAS commissioned, only 23% of Southeast Asian organisations are transforming their operations with AI, reflecting a cautious approach in emerging markets such as Malaysia and Thailand.

“Although the availability of generative tools to consumers made AI seem magical, enterprises must have a solid data foundation and a well-organised methodology in place before they can effectively and responsibly use AI,” says Siboro. “The basis of AI is data. The AI results will be biased if the data is contaminated. For this reason, we are using our data ingestion and preparation skills to assist businesses in producing clean and trustworthy data for AI.”

The most immediate business benefits to expect in 2025 are expected to come from AI applications that enhance operational efficiency and mitigate risks, says Siboro.

Automation, predictive analytics, personalisation and fraud detection are key areas in which organisations can focus their efforts. In addition, leveraging AI to streamline internal processes and accelerate innovation will position businesses for sustained success — with cybersecurity remaining a priority.

Malaysia’s RM50 million investment in AI education highlights a national ambition to cultivate technological leadership, too.

For businesses, this push towards AI adoption presents opportunities to improve productivity, enhance customer experiences and streamline operations. As fully AI-enabled organisations emerge as market leaders across industries, local companies must focus on AI’s practical applications to achieve measurable benefits. One key area in which AI can deliver immediate value is predictive maintenance.

“Using AI to forecast when machinery is likely to break by evaluating both historical and current data from equipment sensors reduces unplanned repairs, minimises downtime and helps schedule maintenance on time,” says Siboro.

As large language models (LLMs) become commoditised, there rises a shift towards domain-specific AI applications that address industry-specific challenges — for instance, healthcare, finance and legal sectors will have much to gain from domain models that are tailored to their unique needs and provide deeper insight and accuracy.

Implementing such models will, however, require high-quality, domain-specific data and substantial computational resources, he says.

Nevertheless, the adoption of Gen AI models is already rapidly growing: an SAS global study showed that 54% of organisations surveyed had begun implementing Gen AI models and 86% planned to invest in the technology in the next financial year.

It is also important to consider the concept of the “Great IT Rationalisation”, as it has long underscored the importance of integrating AI into IT infrastructure to streamline operations and reduce costs.

“Businesses have historically operated using compartmentalised systems, each of which caters to a distinct function or clientele. Because of the burdensome integrations, IT teams are unable to give their businesses the agility they require. Business executives will use the cloud to streamline their vendor relationships and IT infrastructure, gain crucial speed and save expenses as part of the impending Great IT Rationalisation.

“The biggest benefits will accrue to those who modernise on a cloud-native, AI-powered platform that powers several operations. They can accomplish democratised, integrated data and decision-making capabilities that cover the entire organisation and the customer life cycle,” says Siboro, adding that ensuring sustainable AI investments will require businesses to prioritise efficiency and scalability.

“It is impossible to overlook speed and algorithmic effectiveness as crucial levers for lowering cloud usage. AI, which consumes a lot of energy, will continue to push for nuclear and other sustainable energy sources, but it will also raise demand for more energy-efficient versions. We need to improve the efficiency of AI models, just as the automotive and household appliance industries achieved significant strides in energy-efficiency.”

Save by subscribing to us for your print and/or digital copy.

P/S: The Edge is also available on Apple's App Store and Android's Google Play.