At the recent AWS re:Invent conference in Las Vegas, Amazon Web Services (AWS) made several exciting announcements regarding their platform for creating and running AI applications, Amazon Bedrock. AWS chief Matt Garman highlighted the scale of advancements in Bedrock and mentioned that Swami Sivasubramanian, vice president of AI and Data at AWS, would provide more details in his keynote speech the following day. Garman also emphasized that Bedrock is the easiest way to build and scale generative AI applications.
One of the major updates to Bedrock is the addition of Automated Reasoning Checks, a tool designed to prevent AI from making factual mistakes or hallucinations. This feature is particularly useful for industries such as healthcare and finance, where accuracy is crucial. Garman stated that Automated Reasoning Checks do not require advanced AI expertise and can ensure that AI provides correct and trustworthy answers. PwC is already using this tool to build accurate and trustworthy AI assistants and agents for their clients.
Another new feature in Bedrock is Model Distillation, which allows users to shrink large AI models into smaller ones without sacrificing accuracy. This results in faster and cheaper models, with Garman sharing that Model Distillation in Bedrock can deliver models that are 500% faster and 75% cheaper. Robin AI is already using this feature to save money while providing quick and accurate answers for legal questions. Customers can choose the optimal model for their use case and a smaller model from the same family, balancing application latency with cost efficiency.
AWS also showcased Bedrock’s ability to manage and coordinate multiple AI agents for large-scale, complex workflows. This feature is particularly useful for companies like Moody’s, which uses Bedrock’s multi-agent system to improve risk analysis. Each agent handles specific tasks, making their assessments faster and more accurate.
In conclusion, AWS is positioning Bedrock as a leader in the field of generative AI by offering innovative tools and models, as demonstrated by the use cases of PwC and Moody’s. With a commitment to security and privacy, Bedrock has become essential for customers looking to incorporate generative AI into their applications and businesses.
