Amazon Web Services (AWS), a cloud computing provider, is expanding its generative AI services in an aim to make the technology more accessible to cloud-based organizations. The new services include Amazon Bedrock, a set of foundation model AI services. The service contains models developed by AI21, Anthropic, and Stability AI along with ones developed by AWS as a set called Amazon Titan. Amazon EC2 Inf2 cloud instances have been made generally available which are powered by AWS’ own Inferentia2 chips and provide high performance AI. Furthermore, Amazon CodeWhisper generative AI service for code development is now generally available and free for individual developers.
The goal of the Amazon Bedrock is to provide users with a set of foundation models which can be customized and trained with AWS to fit individual needs. These models will be integrated with the rest of AWS’ platform, meaning organizations will have easier access to data stored within Amazon S3 object storage services as well as being able to benefit from AWS’ access control and governance policies. Amazon Titan is designed to provide an alternative model to those AI services already available. At launch, two different flavors of the Amazon Titan model will be made available; a text model for content generation, and an embedding model which creates vector embeddings and can be used for things like complex search capabilities.
As well as expanding AI services, Amazon is also focusing on responsible, explainable AI. Amazon Bedrock will have filters to make sure all data sets are filtered for inappropriate content, and the output from the modules will also have filters to ensure it is appropriate. Furthermore, CodeWhisperer has been improved for lower latency, as well as to provide a reference tracker to ensure proper attribution for code generated by the service. Security validations are also incorporated and the code is checked for potential vulnerabilities. Demonstrating their commitment to the technology and its wider utilisation, Amazon have made CodeWhisper free for developers.
Bratin Saha, VP of ML and AI Services at Amazon explained the push as a desire to “democratize the use of generative AI” and make it available across industries. With its expanded AI Push, AWS is looking to challenge rivals such as Microsoft Azure and Google Cloud, who are also continually rolling out their respective services to benefit organizations of all sizes.
