Despite the buzz around it, generative AI is only as transformative as how it is used.
“Similar to the journey to the cloud, [adopting generative AI] enables a conversation that goes well beyond technology. [This is why] we’re focused on helping our customers transform so that they can get the most out of their data to accelerate growth in a world where disruption happens every day, which can be through improved customer experience or unlocking new business models,” Jaime Valles, vice president for APJ at Amazon Web Services (AWS), tells DigitalEdge on the sidelines of AWS re:Invent 2023 in Las Vegas.
He adds that Amazon Bedrock is one of the ways AWS is democratising access to generative AI. The service allows organisations to build and scale generative AI applications with a selection of foundational models (FMs) by accessing a simple application programming interface, in a secure environment and without managing any infrastructure. The FMs offered include those from AI21 Labs, Anthropic, Cohere, Meta, Stability AI, and Amazon.
Valles expects Amazon Bedrock to be beneficial and in demand by organisations that want to scale and customise generative AI solutions without building them from scratch. The National University Health System (NUHS) in Singapore, for instance, plans to use Amazon Bedrock and generative AI to drive better patient experience. It is experimenting with Amazon Bedrock to develop a solution that automates patient discharge summaries so that its clinicians can focus on their consultation with patients.
Meanwhile, Bangkok Biznews is leveraging the Claude2 foundation model on Amazon Bedrock to fine-tune and develop summaries of their news content, making it quicker and easier for readers to consume content on the go.
As for Malaysia’s 123.rf, its AI Image Generation service uses Stability AI on AWS with significant application of Amazon Sagemaker. This has led to improvements in the user experience and product lineup, as well as licensing rates of AI-generated content to increase by 20% since its launch.
See also: Testing QA New Section BDC Feature Winner 1
“We’re still in the early days of generative AI and there won’t be one model that will address all the needs of a customer so we’re convinced that democratising access to generative AI will [be better] than having a narrow approach. We’re doing so by offering customers choice [of FMs], safety and security. While customers can customise their selected FMs with their data, that data will remain with them [and be completely under their control. No customer data will be used to train or improve the original base model],” asserts Valles.
