Amazon Web Services (AWS) has announced new capabilities for its Bedrock artificial intelligence service, including the ability for customers to import their own custom AI models to run alongside models from major providers. This move is part of a larger effort by Amazon to cater to AI application developers and expand the capabilities of its cloud platform. The Bedrock managed service offers access to a variety of AI models, putting Amazon in competition with other cloud giants like Microsoft and Google, as well as newer AI platforms such as OpenAI and HuggingFace.
The new Custom Model Import feature allows customers who have developed their own AI models outside of Bedrock to bring them into the platform and integrate them as first-class models alongside other models. This feature is available in preview starting today, with early users including Amazon’s internal Rufus team, which is using the feature to share the model that powers its generative AI shopping chatbot with other developers within Amazon. This move aims to make it easier for developers to leverage their existing custom models within the Bedrock ecosystem.
In addition to the Custom Model Import feature, Amazon is also launching updates to its Titan foundation models within the Bedrock service. The Titan Text Embeddings Model V2 will be released next week, improving accuracy and efficiency, while the Titan Image Generator is now generally available and includes built-in invisible watermarking for identifying AI-generated images. Furthermore, third-party models such as Meta’s Llama 3 foundation models and Cohere’s Command R and Command R+ models for enterprise AI applications will soon be available on Bedrock, following the recent release of Anthopic’s Claude 3 Opus model.
AWS is also making its Model Evaluation capability, which helps developers compare different models, and its Guardrails feature, which incorporates safeguards to remove or block personal and sensitive information, profanity, and harmful content, generally available within the Bedrock service. These features aim to provide developers with additional tools and resources to optimize and protect their AI models while using the platform. By offering a range of capabilities and third-party models within Bedrock, Amazon is aiming to attract a wider range of AI application developers and solidify its position in the competitive AI landscape.
