A recent report titled by research firm IDC has revealed that approximately 32% of organizations surveyed in the Asia-Pacific region have expressed their commitment to investing in generative AI technologies.
Generative AI is a type of artificial intelligence system capable of generating text, images, or other media in response to prompts.
Additionally, 38% of respondents are actively exploring use cases for implementing generative AI.
These organizations, which prioritize digital-first strategies, aim to utilize generative AI as a crucial tool to enhance their overall business intelligence and drive efficiency across various functions.
These functions include marketing, sales, customer care, research and development, design, manufacturing, supply chain, and finance.
In the Asia-Pacific region, the foremost use case for generative AI is knowledge management. It is leveraged to facilitate access and search across extensive repositories containing different types of information such as images, documents, voice recordings, and other formats within an enterprise.
The second prominent use case involves code generation, where application programmers adopt generative AI to create, optimize, complete, test, and debug code. This implementation leads to improved programmer productivity and higher quality in the code developed.
Additionally, generative AI finds application in marketing automation and customer-facing roles, such as conversational applications. In these scenarios, marketers can leverage generative AI to generate highly customized marketing content and create search engine-optimized material.
“Generative AI has the potential to reimagine the organizational landscape in a completely new way. However, the
inherent complexities and risks around implementing the same needs to be carefully assessed,” said Deepika Giri, head of research for Big Data & AI at IDC Asia-Pacific including Japan (APJ) Research.
“Generative AI technology is also largely in its early stages, as vendors are unable to fully address the privacy, security, accuracy, copyright, bias and misuse concerns around this groundbreaking technology,” Giri added.
Various stakeholders, including hyperscalers, cloud service providers, AI engineering companies, specialist storage companies, and investment firms, are actively involved in the field of generative AI, according to IDC.
These stakeholders offer Model as a Service (MaaS) solutions, point solutions, infrastructure hosting, and investment opportunities related to this technology.
To support the development and testing of large language models (LLMs), companies are emerging that provide synthetic training data.
This data addresses concerns like the use of sensitive information and bias, considering the vast amount of data required for training.
Practical adoption of generative AI can range from acquiring ready-to-use solutions for marketing, customer care, and code generation, as numerous vendors embed generative AI capabilities in their offerings.
Alternatively, LLMs can be adopted and trained or fine-tuned for specific use cases, which can be resource-intensive in terms of computing power and energy consumption.
A simplified approach called prompt engineering involves using “natural language” queries to train the model effectively.
Another technique, prompt tuning, has emerged, allowing training without retraining the model or adjusting parameters, thus reducing the computational requirements. These approaches aim to strike a balance between extremes.
Regardless of the adoption approach, there is an inherent cost associated with the underlying infrastructure due to the compute-intensive nature of the models, IDC said.
This cost can be either an upfront investment in setting up a data center or included in the price of the MaaS offering.
The application of generative AI raises global concerns, as there is currently limited governance in place.
Regulatory bodies face pressure to address issues concerning data privacy, security, intellectual property rights, and the potential misuse of AI-generated content.
Governments often adhere to existing policies rather than formulating new ones in this evolving landscape, IDC said.
