Published on October 16, 2023, 3:20 pm
The IDC Survey Spotlight report on the adoption and application of Generative AI in Asia/Pacific organizations has revealed some interesting findings. According to the report, 32% of the surveyed organizations in the region are committed to investing in Generative AI technologies. Additionally, 38% of the participants are exploring use cases to implement using Generative AI.
These digital-first enterprises see Generative AI as a valuable tool for elevating enterprise intelligence and driving efficiencies across various functions such as marketing, sales, customer care, research & development, design, manufacturing, supply chain, and finance.
One of the key use cases for Generative AI in Asia/Pacific is knowledge management. Organizations leverage generative AI to access and search large repositories of different types of information across their enterprises, including images, documents, voice recordings, and more.
Another important use case is code generation. Application programmers adopt generative AI to create, optimize, complete, test, and debug code. This leads to improved productivity and quality of code developed.
Generative AI also finds applications in marketing automation and customer-facing roles. Marketers can use generative AI not only to generate highly customized marketing content but also to create search engine-optimized content.
Despite its potential to reshape organizations in new ways, implementing Generative AI comes with inherent complexities and risks that need careful assessment. Deepika Giri, Head of Research at IDC Asia/Pacific including Japan (APJ) Research points out that vendors still face challenges around privacy concerns, security issues, accuracy limitations, copyright infringement risks, bias concerns, and misuse possibilities related to this groundbreaking technology.
With the growing interest in Generative AI adoption and application comes a wave of vendors competing for opportunities in this space. These include hyperscalers and cloud service providers offering Model As A Service (MaaS) offerings, specialized storage companies selling infrastructure for hosting generative AI solutions ,and investment firms seeking substantial returns from betting on this technology.
However, the adoption of Generative AI can range from procuring ready-to-use solutions for marketing, customer care, and code generation to adopting and training large language models (LLMs) for specific use cases. Training LLMs requires significant computing power and energy, which can be costly. Prompt engineering has emerged as a simplified way to train these models by using natural language queries to elicit the desired responses.
An emerging technique called prompt tuning offers a balance between retraining models from scratch and tweaking parameters. It provides a simpler way to train the model without incurring extensive compute requirements.
It’s important to note that leveraging Generative AI comes with underlying infrastructure costs since the models are compute-intensive. These costs can take the form of upfront investments in data centers or pricing included in MaaS offerings.
While interest in Generative AI grows, there are growing concerns globally about its application. Regulatory bodies face pressure to address issues related to data privacy and security, intellectual property rights, and potential misuse of AI-generated content. However, there is currently no consistent legislation for generative AI in Asia/Pacific countries. Each country is evolving its approach to AI regulation at various stages, aiming to strike a balance between promoting innovation and addressing regulatory concerns.
In conclusion, the IDC Survey Spotlight report highlights the increasing interest in Generative AI adoption and application among Asia/Pacific organizations. The potential benefits across various functions are driving this enthusiasm. However, organizations need to carefully assess the complexities and risks associated with implementing Generative AI technologies before fully embracing them.
(Note: This article does not promote any specific original source mentioned in the source article.)