Generative AI has gained significant traction in the tech industry across Asia, with around 72% of companies exploring or investing in it by 2023. Japan leads the region in AI adoption, followed by Australia, India, and Singapore, where governments have made substantial investments in AI initiatives. To delve deeper into the implications of generative AI, we spoke with Deepika Giri, Head of Research at IDC APJ.

Deepika oversees BDA, AI, blockchain, and Web3 research programmes across Asia/Pacific. A  seasoned data and AI professional, Deepika has over 20 years of experience in IT services and leadership positions at Capgemini, Infosys, and Accenture. 

Here are the edited excerpts.

How would you describe the current state of artificial intelligence (AI) adoption and its impact across various industries globally?

The adoption of AI has experienced significant growth across multiple industries globally, including in the Asia/Pacific region. Conversational AI, such as AI chatbots, has seen increased usage, along with computer vision applications and text-to-speech technologies. The overall software market is projected to have a compound annual growth rate (CAGR) of 11.2% between 2021 and 2026. However, when considering AI platforms alone, the growth rate is even higher at 40.4% over the next four years, making it one of the fastest-growing categories in the technology landscape.

Several primary use cases have emerged as rapid adopters of AI. These include enhancing customer experiences, implementing business process automation and intelligent process automation (IPA), and industries like banking, financial services, and manufacturing have particularly embraced AI investments. Cloud adoption has played a crucial role in the adoption of AI, with AI workloads on the cloud expected to grow at a CAGR of 51.2%. Currently, approximately 56% of AI workloads are on the cloud, but this is projected to reach 73% by 2026.

And how does it look in India and Singapore?

From a market perspective in the Asia Pacific region, Japan stands as the largest and most mature market for AI adoption followed by Australia, India, and Singapore.  These markets have shown significant focus on AI adoption, particularly in the public sector. The governments in these countries have made substantial investments in AI, leading to the development of automated citizen services and initiatives related to threat intelligence and defense. Singapore, in particular, has been proactive in driving AI advancements.

In India, the focus is more on providing communities, especially those in remote areas, with access to information. The government has recognised the potential of AI, particularly with the imminent launch of 5G technology. 

What are some of the other notable trends that you see in the field of generative AI?

Our recent report revealed that around 72% of tech companies in Asia were either exploring generative AI or had already invested in it by 2023. This indicates a substantial presence of generative AI technologies in the region.

In terms of use cases, one of the areas with the highest adoption is knowledge management. Organisations are leveraging generative AI to sift through vast amounts of information in various formats and extract valuable insights. Another notable use case is coding, where generative AI is used to enhance developers' coding productivity. Additionally, there is a growing trend toward automating the generation of marketing content and campaigns using generative AI.

The generative AI space is expanding beyond the major hyperscalers like Google and Microsoft, with numerous players entering the market. Companies specialising in engineering and developing models, fine-tuning them for specific use cases, are in high demand. There is also a need for processors and coprocessors to handle the computational requirements of these models. Cloud service providers play a crucial role by offering infrastructure and models as a service, allowing users to access and utilise trained models for inference purposes. Storage solutions are also essential for storing the infrastructure, whether it's provided by cloud service providers or private data centers.

Large companies such as Microsoft are embedding generative AI capabilities throughout their software stacks. It is worth noting that instead of trying to cover all possibilities, the trend is towards targeted point solutions. This allows smaller companies, such as marketing firms, to leverage AI without extensive infrastructure investments. Not all use cases require large-scale language models with billions of parameters. Smaller, medium-sized models can be equally effective for specific use cases.

Can you provide examples of successful implementations of generative AI, or is it too early?

There have been significant advancements in the field of generative AI. The adoption of large language models and embedded analytics is becoming mainstream. The focus has shifted to transformer models, which can generate content based on existing information. Companies like Microsoft, OpenAI, and Google have been leading the way by investing in research and development for the past five years.

However, the game-changer was the introduction of models like ChatGPT, which made AI more accessible to end-users. Previously, AI adoption was limited to banks and large businesses, but models like ChatGPT democratised access to AI and simplified its complexity. Success stories include applications in securities management, workspace management, crowd control, and information dissemination, enabling increased accessibility and availability of information.

While generative AI is still evolving, it has already shown promising results and has the potential to drive further advancements in various industries.

How do you view the larger issue of regulation in the context of generative AI, particularly when it comes to ensuring accountability and ethical use?

When organisations adopt generative AI technologies, they must ensure that the information generated and published is not detrimental to the organisation or its stakeholders. Having humans involved in the process will continue to be essential to protect against ethical concerns and confabulation.

Another challenge is the validation and credibility of data sources. As data is fed into these models from various sources, it becomes necessary to establish guardians and mechanisms to verify the credibility of the data. This ensures that the information consumed at the end is reliable and aligned with the desired outcome.

Generative AI is largely unregulated because there is still a lack of understanding about how to create effective guardrails and regulations. Many countries and organisations are in the process of exploring the technology. While they are open to adopting generative AI and harnessing its potential, they are taking a step-by-step approach. This includes evaluating specific use cases, assessing potential downsides, and implementing appropriate governance measures. The goal is to introduce generative AI solutions in the market while ensuring they are thoroughly tested and do not pose any significant risks.

Regulation in the field of generative AI is an ongoing discussion, and as the technology continues to evolve, we can expect to see increased focus on addressing ethical concerns.

How do you see the impact of generative AI on employment? While it is expected to transform work dynamics, do you believe it will also generate new opportunities instead of solely replacing jobs?

The economic impact of generative AI and job loss is a significant concern. However, it is important to understand that the introduction of AI does not necessarily eliminate jobs completely. Instead, it has the potential to transform the nature of work.

Generative AI is not about eliminating jobs but rather about increasing efficiency and productivity. Just as automation technologies in the past, such as ERP systems, changed the way we work without eliminating jobs, generative AI is expected to bring similar transformations.

While there may be concerns and uncertainties during this technological shift, it is crucial to understand that new technologies often lead to the emergence of new economic factors and revenue channels. It is a transformative stage in technology, and although it may seem threatening, history has shown that advancements tend to enhance our lives and create new avenues for economic growth.

Let's delve into the practical aspects of adopting AI. Is there a recommended playbook or set of questions that leaders should consider when embracing AI?

When embarking on the AI journey, leaders should begin by defining their goals and understanding what they aim to achieve with AI implementation. 

• Identify potential risks and be aware of the ethical implications associated with AI technologies. 
• Hire the right talent, which can be a challenge. Actively build expertise and knowledge in AI.
• Start with low-risk use cases that can provide tangible value in areas like IT operations, knowledge management, or customer care. Incremental and iterative implementation often yields better results compared to a large-scale deployment strategy.
• Find trusted partners, such as software vendors or system integration partners, to deploy and manage complex AI technologies. 
• Select the right use cases that align with the organisation's business requirements and ensure that there is executive-level involvement. Their understanding of how AI can benefit their respective areas, such as finance or production, is instrumental in gaining support and buy-in.
• Establish a center of excellence or AI labs and sandboxes within the organisation to foster a culture of experimentation and learning. 
• Educate the board and other business stakeholders about the potential and limitations of AI. CIOs play a role in setting realistic expectations, introducing AI initiatives gradually, and garnering support from top-level executives.

HR leaders have a vital role in managing the impact of AI on job roles and facilitating change management. They should educate employees about the initiatives being implemented, clearly communicate how AI will augment and enhance their work, and address any concerns or misconceptions. Change management is often underestimated but is pivotal in ensuring the successful adoption and acceptance of AI initiatives.

What do you predict for the future of generative AI? What are the major challenges ahead?

Over the next two years, generative AI will witness rapid growth and increased interest across domains. Businesses will prioritise quick and easy implementation of solutions, with ready-to-use options gaining prominence in areas like marketing, customer care, and code generation. We're already seeing AI capabilities integrated into existing enterprise applications for seamless data connectivity and automation.

To fully leverage generative AI in the enterprise, organisations need the right tools and partnerships. The challenge lies in connecting silos and developing a common architecture that spans verticals and functions. Collaboration among technology vendors to establish standards and interoperability will be crucial.

Time to value is critical for generative AI adoption. Vendors must deliver solutions that enable efficient workflows, allowing organisations to quickly derive value. Establishing an open standard and collaborative data ecosystem will facilitate easier adoption and shorten the time to value.

Data security and compliance remain significant concerns. Vendors must address regulations like GDPR in the EU or local requirements in ASEAN countries. Compliance ensures responsible use and builds trust in generative AI.

What aspect of generative AI excites you the most, and what concerns you when considering its implications?

AI’s potential to positively impact developing economies, such as India. AI can reach remote rural areas and improve agriculture, education, and access to valuable information, fostering inclusive growth and development.

On the other hand, I am concerned about the misuse of information and the risks associated with AI models. Responsible and ethical use of AI, along with robust data governance practices, is crucial to mitigate the risks and ensure the effectiveness of AI.