AWS Bedrock

Amazon Bedrock is a fully managed service providing access to various foundational AI models from providers like AI21 Labs, Anthropic, Cohere, Meta, Mistral AI, Stability AI, and Amazon. Bedrock is suitable for developing AI applications, offering flexibility in model selection and robust security features, similar to Azure AI.

Here’s what Amazon Bedrock offers:

  • Choice of Foundation Models: Access to an array of models from top AI providers like AI21 Labs, Anthropic, Cohere, Meta, Mistral AI, Stability AI, and Amazon. This variety allows you to select the model that best fits your specific needs.

  • Generative AI Applications: The service is geared towards building generative AI applications, which can generate text, images, or other data types based on your inputs.

  • Customization Options: Customize models to better suit your specific requirements. Techniques such as fine-tuning, where the model is adjusted to perform well on your particular data, and Retrieval Augmented Generation (RAG), which combines information retrieval with generative capabilities, are available.

  • Security and Privacy: The service ensures that all interactions with the AI models are secure, and your data remains private, which is crucial for maintaining trust and compliance in business applications.

Amazon Bedrock is comparable to Azure AI and is ideal for developing AI applications where you want the possibility to experiment with different models to find the best fit or looking to deploy a robust AI-driven application. 

Details

Availability: This tool is available to staff and faculty at Yale with an AWS account. 

Integration Capabilities: AWS Bedrock integrates in multiple ways including AWS services and API access. 

Pricing: Depends on the usage. 

Data Risk Classification: Approved for high risk data, however, it needs to undergo an AI request review process. 

Privacy and Protection: The tool includes privacy and protection measures to ensure data security. 

How to Access

Please note that to use this service, you need to get your use case approved by Data Governance. You must follow the AI request review process and use this form to request access to AWS Bedrock. For these types of requests, please follow the steps outlined as you can’t submit the request form in ServiceNow before your request has been approved by Data Governance and the ISO team. For more information, check out the Amazon Bedrock documentation.