To say generative artificial intelligence (GenAI) took centre stage in 2023 would be an understatement. Early in the year, industry players were wildly enthusiastic about the technology’s potential, but things quickly took a cautious turn after the public release of OpenAI’s large language model (LLM) programme, ChatGPT.
The pervasive utility of this hyper-productivity tool in automating mundane tasks not only made headlines but also inspired a string of rival chatbots — Google’s Bard, Anthropic’s Claude and Elon Musk creation, Grok.
Amid the excitement, Open AI’s capabilities underwent rapid enhancement, transitioning from the GPT-3 LLM to GPT-3.5 in January and to the subscription-only GPT-4 in less than a year. The swift pace of its growth unsettled regulatory bodies worldwide, prompting governments to urgently address the potential hazards associated with GenAI.
Concerns over issues such as job losses, copyright infringement and what is fake or not surfaced. In March, top tech pundits signed an open letter, urging industry players to exercise restraint in the development of AI for the future of humanity.
While the letter may have appeared melodramatic initially, its impact was palpable. Before long, industries and governments responded by taking proactive measures, enacting policies and guidelines in the hope of fostering the ethical and safe use of AI technologies.
Contrary to expectations, this collective response did not hinder the development of GenAI; rather, it accelerated its progress. Experts across industries, cautiously optimistic of the benefits of AI, could not ignore its economic potential.
In September, Malaysia’s Deputy Economy Minister Datuk Hanifah Hajar Taib said GenAI was estimated to unlock US$113.4 billion (RM533.5 billion) in productive capacity, propelling Malaysia into becoming a high-income nation. This is equivalent to a quarter of the country’s gross domestic product in 2022.
As the year comes to a close, Digital Edge speaks to Minister of Science, Technology and Innovation (Mosti) Chang Lih Kang; Fusionex director of customer engagement, AI and analytics, Freddy Loo; and GXBank CEO Lai Pei-Si to gain insight into what we can expect in the coming year.
(Photo by GXBank)
Digital Edge: Generative AI took centre stage in 2023. How has it affected the country’s socioeconomic landscape and how do you see it affecting the year ahead?
Freddy Loo: The last 12 months have been very exciting with the introduction of ChatGPT-3, and since then [it] has sparked a flurry of investments and a race by the largest tech companies to out-innovate each other. In Malaysia, the man on the street, even those without any AI knowledge, had unprecedented access to its power through a consumer-friendly interface. You can imagine how exciting this is, almost like KITT from Knight Rider!
On the economic front, GenAI can unlock cross-sectoral-wide gains in productive capacity. Naturally, manufacturing and services will have large gains since it’s one of the largest contributors to Malaysia’s economy and comprises a large workforce.
Creative industries like media and entertainment would receive a boost from the ease of content generation — copywriting, images and videos. Obviously, this introduces other ethical issues such as rights attribution to the owner of the training data, but we will leave that for another day.
The biggest sectors [affected] would be journalism and art and culture. GenAI may undermine news credibility and trustworthiness, and can produce fake or misleading news to influence public opinion. For the latter, it may diminish the value of original artwork as it can replicate the styles of human artists and creators. It can also create much more diverse content to cater for different perspectives, better than humans can.
Lai Pei-Si: Some elements of GenAI are entrenched in a lot of the tech we already use, from the hyper-personalised offers that e-commerce platforms are surfacing out to consumers, to the workforce planning proposals offered to businesses.
(Photo by Zahid Izzani/The Edge)
We are now on the cusp of seeing an even more far-reaching impact of GenAI in the way we do business and run our lives every day.
The biggest gainers are industries that have copious amounts of data and have the ability to store, arrange and organise it by leveraging technology and automation capabilities. Clear winners are companies that are set up to use technology in their core business [and] have access to platforms that can make sense of data collected and optimise its respective use cases.
For proper GenAI adoption to take place, data privacy and security due consideration, including ethical use, is necessary. Therefore, industries like the financial sector would be best placed for its burgeoning development.
As a digital bank, we will benefit greatly. This begins with ensuring financial access to underserved communities — not just individuals, but local businesses as well — who struggle with access to the financial support they need, or find traditional banking and its processes a challenge.
That means leveraging the wealth of data we hope to gather from our customers and partners, which will give us strong insights and understanding of their financial needs and behaviour.
The biggest concern is job losses. Looking at history and how jobs changed with technology, how do you see the current job market evolving in the coming year and can AI displace people?
(Photo Fusionex)
Chang Lih Kang: Job evolution is inevitable with the rapid growth of GenAI. This does not mean that AI will replace people entirely in the job market. It is more likely that the inculcation of GenAI will change a job’s focus, allowing people to leverage AI tools to enhance existing work tasks for more effective and efficient results.
The current job market will evolve in the coming year due to the impact of GenAI. Some foreseen trends are the automation of routine tasks, creation of new job roles, increased demand for soft skills, continuation of the upskilling and reskilling programme and the rise of the gig economy.
There will be a need to ensure we have the right skills and training. This is part of what our education and skilling infrastructure will need to help address. The key point is that the existing workforce must equip themselves with the latest skills, knowledge and work in tandem with the technology update to enable them to take up the new role.
Lai: GenAI will redefine the job landscape, creating new types of jobs, and more of it. History has shown us many times that while technological advancements impact the job market, they have also expanded the choice of careers — always finding new ways to stretch their boundaries.
At GXBank, we have expansive roles that ride on the amazing wave of new technologies — roles like data science modellers, data analysts, security specialists and UI/UX teams were created. While team members fill up more “traditional” roles like customer service and operations, risks and product, are all expected to be technologists as well. We also recognise that to create the best digital banking experience, it requires a personal touch and creativity that can never be replaced by AI.
In March, we saw tech leaders and top executives calling for the industry to pull the brakes on AI development as it may be detrimental to mankind. Some called it a tall other, others said there was some truth in it. What are your thoughts?
Loo: The GenAI wave is real and it’s coming whether we like it or not because the technology’s access is so easy. With that power, actors of harm can easily utilise it for malicious intent, but [the solution is] not as simple as authorities putting in guardrails.
Guardrails would impede AI development and shift its development from legit actors to deceitful actors who will operate with no bounds. The shift [will also be to] countries that operate with fewer controls and less-than-noble intentions.
From another angle, guardrails could also discriminate against smaller organisations that have less resources and capability to leverage technology for their economic needs. Larger organisations likely will thrive and create this bubble of “creative consolidation”.
I think it’s important to form guidelines based on a risk-based approach, sectors and cross-border operability. Regulations also need to be reviewed and adapted since AI is moving at the speed of light. Regulations developed today might be irrelevant very quickly.
Chang: The ongoing debate reflects a nuanced landscape of both promise and peril. Concerns persist regarding its ethical implications, potential job displacement and the need for robust regulation.
Regulators need to be alert to high-risk uses of AI and it is in this context that the concept of brakes becomes critical. The good news is that many sectors that can pose a high risk or that constitute critical infrastructure are regulated sectors. This means the governance infrastructure needed to manage risks are already in place to a large extent.
Striking a balance between harnessing the benefits of AI and addressing its societal challenges remains a crucial task. [This requires a] concerted effort in ethical development, regulatory frameworks and ongoing dialogue to ensure that AI contributes responsibly to the betterment of mankind.
Calls for AI regulations are loud but until the government puts in a proper governance framework, how can people self-regulate to prevent AI from potentially being harmful?
Chang: Any AI output must be assessed by a human user before it is used. Key baseline practices, such as not freely sharing personal information and being disciplined in checking an information source’s credibility, remain the same.
While waiting for the AI governance and code of ethics to be ready in 2024, the protection against harmful AI elements is covered under the Communications and Multimedia Act 1998, under the purview of the Malaysian Communications and Multimedia Commission. The act is technologically agnostic and works vertically across technologies.
Key sectors where risk may be material already have a governance framework in place, such as in financial services. In the meantime, institutions and companies should be free to create guidance for themselves based on the risks they see arising.
While guidance on the safe use of AI is important, actual practice in responsible user behaviour is even more critical. Just as the best solution to mitigate the harms of scams and fraud is for everyone to not share information freely or believe in non-credible communication, it will be important for every person to be educated on the correct use of GenAI.
Lai: For us, Bank Negara Malaysia’s Financial Sector Blueprint 2022-2026 provides the guardrails to leverage the best of technology to achieve our goal of building a financially inclusive and resilient Malaysia. We are working with policymakers, regulators and other relevant players in the industry to encourage harnessing the power of AI while balancing the ethical use of it, such as the appropriate parameters of cybersecurity and data protection, among others.
Additionally, our parent companies, GXS and Grab, also have live policies surrounding the use of AI and LLMs, which will continue to be improvised as the use of GenAI matures.
With technology, we can incorporate specialised algorithms that detect and correct potential biases. These will actively monitor for any disproportionate impact on certain groups, allowing us to adjust our processes accordingly.
We are expecting a host of challenges as GenAI continues to make its mark in the world. What would be a massive hurdle and how can it be overcome?
Loo: The one big challenge is that the world’s AI superpowers are locked in a gladiatorial battle for AI supremacy. They recognise the competitive advantage it would have for many years to come across many industries such as defence, commercial and education. An equivalent competition is also happening at the level of the largest tech organisations: Google, Amazon, Meta, Baidu, Alibaba and Tencent.
Competition is healthy, until one uses unscrupulous tactics to inhibit the progress of others, rushes to out-innovate others at the expense of proper testing and controls and overemphasises short-term goals, losing the larger potential of GenAI in addressing global humanity concerns.
This is the geopolitical tension that is happening as we speak while the world is still grappling with climate change, inequality, food security and many more issues. We could be losing precious time in trying to exert and win political and economic dominance.
Chang: One of the challenges is the potential to create biased or discriminatory content. The National ArtificiaI Intelligence Roadmap 2021-2025 has outlined seven responsible AI principles used as a basis in developing the AI governance and code of ethics, which will be the foundation for AI legislation. One of the principles is fairness.
To quote the National AI Roadmap: “Fairness is the foundation for treating people with dignity and respect. If AI systems guide medical treatment, loan applications or employment for example, they should make the same recommendations to everyone with similar symptoms, financial circumstances or professional qualifications.”
The national policy and regulatory framework should be set to guide responsible AI use and should remain in line with global best practices and focus on ensuring digital safety.
Micro, small and medium enterprises (MSMEs) are the backbone of the Malaysian economy. Will GenAI be the bridge that MSMEs are looking for to catch up with the rest of the world or will it widen the gap even more?
Loo: AI development has reduced the barriers of entry in terms of usage complexity, skills required for development and cost. Case in point: The ease of [using] ChatGPT was one of the key reasons that led to the explosion of its adoption. It’s as simple as logging into a browser and asking questions, just like another fellow human being. No programming knowledge and AI knowledge required at all.
This implies that SMEs that typically don’t have the resources or capabilities compared to larger organisations can employ this technology to improve their customer servicing capabilities, streamline operational planning, create compelling artwork for marketing or even improve logistics planning with AI.
SMEs likely would need to work with partners to leverage the solution, rather than build their own AI solution. There are many software-as-a-service solutions that have much lower barriers to entry (that is, cost) and tap a host of proven use cases.
While the lower barriers to entry for AI open up possibilities, it will not be sufficient for SMEs to leapfrog the economy without taking into consideration the factors mentioned above. The fundamental issue of entrepreneurship, talent, supporting infrastructure, incentives and familiarity with the digital ecosystem all contribute to a growing chasm of digital divide.
Lai: The potential is definitely there. The scale of GenAI’s impact on MSMEs will depend on how well they leverage its power and overcome challenges such as access, skills, relevance and integration, and regulatory support.
However, from a banking perspective, the fact that we will be able to combine the data of our customers with AI, the transformative impact of just enhancing internal efficiencies and building customised credit models, is vast and already exciting!
We will be able to generate challenging and complex financial and non-financial models quicker, and continuously evaluate dynamic customer segments and their potential financial behaviours, offering specific insights into potential financial standing.
In the past, there has been limited usage of such alternative data (that is, Natural Language Processing or NLP datasets, geospatial, behavioural, feature usage, navigation), automated feature engineering and advanced complex models such as neural networks and deep learning models as these sophisticated and complex algorithms or techniques are difficult to interpret, explain and implement.
Strategic planning and a proactive approach to technology adoption (whether it is within the business itself or engaging vendors) will help level the opportunities many MSMEs are currently denied due to the lack of access and the complexities of the current banking processes.
In Budget 2024, an allocation of RM20 million was announced to provide for the establishment of the country’s first AI study centre, the Faculty of Artificial Intelligence at Universiti Teknologi Malaysia. Is this enough, and should there be an allocation/incentive for AI upskilling and reskilling in the current workforce?
Loo: Resources for education and research is almost always a good, prudent use of resources as it lays the foundation for societal and economic development. The funding is intended to intensify the exploration of cross-disciplinary AI. I would assume that would create more industry-specific use cases, skilling more relevant employees to join the workforce.
This is great to develop the next generation of AI-savvy workforce. This workforce would be the bedrock in the next five to 10 years. It will take a bit of time for these young talents to gain work experience and industry expertise to be experts in their own right.
From now until then, the country still needs to meet existing demands across different industries. It’s best to also tap the current mid-level workforce and relook its fungibility, leveraging the current domain experience, but amplifying their capabilities with digital skills.
Chang: The terms ‘enough’ or ‘insufficient’ are subjective. This must be looked at from all possible angles such as current economic growth, quality of education and the amount of support from academics and research.
One of the benefits of the AI faculty is producing the required experts for the industry and further supporting the government in transforming the country into an AI nation. To ensure a programme’s sustainability, a suitable funding mechanism has to be planned thoroughly to avoid a financial burden to the government or the public institutions.
For the upskilling and reskilling in the current workforce, the government is looking into some allocation, but this will be discussed thoroughly with respective agencies such as Malaysia Digital Economy Corporation (MDEC) and MyDigital Corporation.
Will GenAI bring in more foreign direct investment/domestic direct investment to grow the economy?
Loo: Definitely. GenAI will be an added advantage to unleash creativity, improve the efficiency of the local workforce and ensure the workforce moves up the value chain. It’s different from previous technological advancements, where the value stems from automation and efficiency. It also has the ability to take on much more complex cognitive tasks.
With that, it can propel a labour market from a low-cost to a high-value, high-creativity workforce that can compete in industries like game development, green technology, high value manufacturing and entertainment. This diversification of skills will be enticing for foreign investors.
The Malaysian government has also been very supportive with AI development and other emerging technologies such as green technology. Coupled with investment in infrastructure and supporting incentives, I do foresee an increase in regional competitiveness and economic diversification, which will bring in FDIs.
Chang: A growing GenAI market contributes to the overall development of the technology ecosystem. The rapid growth of the GenAI market creates new business opportunities.
Foreign and domestic investors may see potential in supporting or establishing companies in GenAI development, applications or related services. Investors may also be interested in supporting research institutions, universities or start-ups engaged in cutting-edge AI technologies.
Countries with a thriving GenAI market may become innovation and research hubs in the AI sector. Countries that establish themselves as leaders in GenAI can enhance global competitiveness. Investors may be drawn to regions with a strong reputation for AI expertise, innovation and a supportive business environment.
AI has become a concern because it can and has become a cybersecurity threat. Your view?
Loo: AI is a double-edged sword. It benefits cybersecurity as well as lowers the barriers to entry for malicious actors. This technology is just like any other, such as nuclear. As devastating as technology can be, if given proper development, controls and international cooperation, it can be harnessed for good.
AI and nuclear have more similarities than we can imagine. Both have a wide-reaching, intergenerational impact for both good and bad. Countries will and perhaps have already weaponised this technology. All it takes is one rogue actor and it will start a chain reaction of escalating conflicts among superpowers.
However, eight decades of experience managing international nuclear relations can offer some guidance to world leaders in managing this form of technology. The core of nuclear strategy was based on mutually agreed self-control, understanding that the failure of which will lead to mutually assured destruction. Nuclear superpowers formed an alliance and agreement that provide a set of tacit constraints and that are still being upheld till today.
AI should do the same. The international community must agree to a set of guidelines on the use of this technology for the benefit of humankind. What we see now are AI superpowers like the US and China locked in a gladiatorial battle for innovation supremacy that comes at a cost to humankind.
To make matters worse, nuclear is much harder to create but AI has much lower barriers to entry, [making it] accessible to many. Nuclear takes years of research but AI can be deployed in a matter of months or weeks.
We need to move much faster. The world needs more cooperation, not competition, because not every actor in the world will play by the rules. Rules that apply to one country don’t apply to another. We have had 78 years of peace and enjoyed the benefits of nuclear power. Now, we need to strive for the same for AI.
Lai: AI presents both opportunities and challenges in the realm of cybersecurity. The posture you take (offensive or defensive) determines the measures you will take. Regulations and ethical guidelines for the development and use of AI in cybersecurity are essential to mitigate potential risks.
As a digital-only bank, cybersecurity is at the core of our build. Technology and security will constantly evolve, and so will cybercriminals. Our approach to technology and cybersecurity is equally ever-evolving, which means that we will continuously innovate and implement best practice standards to serve our customers efficiently and securely.
