Generative AI revolution powered by Asia-Pacific startups

March 12, 2024 | 17:15
(0) user say
Startups have consistently been at the cutting edge of innovation, and in the age of generative AI, they are poised to harness this technology to transform customer experiences and the way we work.

Investors recognise such value and potential and have poured more than $21 billion into generative AI startups since the start of this year through September 30 2023, up from $5 billion in 2022, according to PitchBook data.

Generative AI revolution powered by Asia-Pacific startups
Gaurav Arora, director and head of Startup Business in Asia Pacific and Japan at Amazon Web Services (AWS)

Today, more than 5,000 generative AI startups are building their solutions on AWS. These scrappy but nimble innovators are disrupting industries with new ideas, developing locally relevant solutions, and introducing new ways of using AI. The achievements of these startups are commendable given the challenge of navigating a rapidly evolving technological space.

Generative AI has captured the world’s attention for its ability to learn and apply knowledge – powered by foundation models (FMs) that are pre-trained on vast volumes of data. However, models are only as good as the data they are trained on. For instance, when Large Language Models (LLMs) pre-trained in English are tasked with non-English queries, they can produce errors and misinterpretations.

This is particularly important to the Asia-Pacific region, where the population speaks about 2,300 languages. To cater to the region’s diverse ways of working, cultures, and languages, there is an urgent need to train LLMs on culturally diverse data to build a more nuanced understanding of human experiences and complex societal challenges. The creation of more culturally aware and localised AI will increase the accessibility of AI technology, impacting countries, communities, and generations to come.

Startups are leading the way in training models with data that represents local text, imagery, audio, video, and other datasets. One such example is Stockmark, a Japanese startup that leverages AWS's infrastructure to train Stockmark-13B, an LLM using 13 billion parameters that was trained on over 220 billion tokens of Japanese text, one of the largest datasets of its kind. To develop culturally aware AI, AWS has invested $6 million in Japan and $5 million in South Korea to help startups, enterprises, and developers build FMs and LLMs trained on local data.

Startups play an important role across the entire spectrum of generative AI innovation, ranging from establishing the foundational frameworks with FMs to developing practical real-world applications. To date, startups using generative AI have already transformed a wide variety of industries like healthcare, financial services, media and entertainment, education, and gaming. By automating tasks, enhancing decision-making processes, and personalising user experiences, generative AI continues to revolutionize how businesses and organisations of all sizes operate.

In the medical arena, AI-assisted training tools can create personalised and real-time training for medical professionals at scale to help address the current talent crunch. Australia-based startup SimConverse uses generative AI in role-play training for over 300,000 healthcare professionals in over 150 organisations in Australia and New Zealand.

The company has used Amazon compute services to train their AI models on over 1,000 scenarios, ranging from simple communication tasks like basic history-taking to linguistically complex de-escalation and the delivery of bad news.

Another industry evolving with AI is the media and entertainment sector, where companies are using generative AI to automate content production, reduce costs, and increase output. South Korea-based startup Toonsquare is pioneering innovations in the webtoon industry with their novel AI-driven generative webtoon production tool, Tooning, powered by AWS. With Tooning, tasks that previously took 60 hours in a week can now be finished in just six hours.

Startups have undeniably played a pioneering role in propelling the generative AI technology revolution from its inception to adoption across industries. As startups continue to drive innovation, they are also facing significant hurdles that encompass resource limitations, ethical and regulatory complexities, integration obstacles, and the absence of in-house expertise or technical proficiency.

To overcome these hurdles, startups need access to cloud resources that level the playing field. AWS recently unveiled technology enhancements at all layers of the generative AI “stack” to make it even easier, cheaper, and faster for companies to build, train, and scale their generative AI innovations.

At the infrastructure (bottom) layer, this includes pushing the envelope on price performance with the next generation in AWS-designed ML chips, such as Trainium2, and introducing new capabilities to accelerate model training and inference with Amazon SageMaker.

For startups looking to experiment with existing models at the tools (middle) layer, Amazon Bedrock offers an expanded choice of leading models, customisation features, agent capabilities, and enterprise-grade security and privacy in a fully managed experience.

At the application (top) layer, startups can transform how they work through applications like Amazon Q, a new type of generative AI-powered assistant that is specifically for work and can be tailored to a customer’s business. Another application is Amazon CodeWhisperer, an AI-powered productivity tool for the integrated development environment and command line. AWS is also helping to plug the AI skills gaps for startups by providing free AI training programmes across the region.

While building is the first step to the generative AI revolution, going global is equally critical in driving widespread adoption of startup-driven generative AI adoption across industries. India-based startup Yellow. AI, a leading conversational AI solutions provider, has listed its generative AI customer service solution on AWS Marketplace, a curated digital catalogue with over 330,000 active customer subscriptions. Today, India-based Yellow.AI handles over 12 billion conversations across more than 85 countries annually.

As startups in Asia-Pacific continue to prove what’s possible by building, customising, and scaling their own FMs or applications in generative AI, they will need access to the most price-performant and cost-efficient ML tools, support, infrastructure, and programmes like AWS Activate to build culturally aware AI. We are excited to support these startups, which will help accelerate generative AI adoption and realise an AI-assisted future world, with fresh perspectives and inventive solutions that will drive positive impact for all.

AWS re:Invent 2023 showcased future of AI AWS re:Invent 2023 showcased future of AI

AWS re:Invent 2023 made a big impression with some groundbreaking announcements, thought-provoking keynote speeches, and immersive experiences to help businesses plan the future of their cloud transformation journey.

AI Singapore brings inclusive generative AI models to Southeast Asia AI Singapore brings inclusive generative AI models to Southeast Asia

Anticipating the importance of inclusive Generative AI models, AI Singapore is collaborating with Amazon Web Services (AWS) to produce the region's first Large Language Model family trained for Southeast Asian languages and cultures.

VIB drives digital banking innovation with Temenos’ latest version of banking platform powered by AWS VIB drives digital banking innovation with Temenos’ latest version of banking platform powered by AWS

Vietnam International Bank (VIB) is delighted to announce the implementation of the Temenos Core Banking solution on cloud this week in collaboration with Amazon Web Services (AWS) and ITSS.

By Gaurav Arora

What the stars mean:

★ Poor ★ ★ Promising ★★★ Good ★★★★ Very good ★★★★★ Exceptional