Amazon is expanding its range of AI services as CEO of Amazon Web Services (AWS), Adam Selipsky, announces a broad range of new capabilities at the annual AWS re:Invent 2023 conference in Las Vegas. With a focus on advancing their signature platform, Selipsky detailed the progress made by AWS in the past year, particularly in the field of AI.
The New Amazon Q: Revolutionizing AI Automation
The highlight of the announcements made by Selipsky was the introduction of Amazon Q, an AI system that will bring generative AI automation to AWS cloud services. This includes development, analytics, and operations. Amazon Q aims to save customers valuable time in architecting, troubleshooting, and optimizing workloads. Selipsky believes that Amazon Q will be an invaluable assistant to developers by answering questions, generating text and visualizations, and taking actions on their behalf.
“We expect that Amazon Q is going to save customers so much time architecting and troubleshooting and optimizing workloads,” Selipsky said.
Additionally, Amazon Q will deeply integrate with multiple cloud services, such as Amazon CodeWhisper and Amazon Connect, enhancing functionality and providing assistance to developers and contact center agents.
Enhancements to Amazon Bedrock: Tailored AI Models and Improved Safety
Selipsky also highlighted several incremental updates to the Amazon Bedrock generative AI model service, which became generally available in September. These updates include new customization capabilities that allow customers to further tailor AI models to their specific needs and data. The customization features enable fine-tuning of models using labeled examples from a customer’s own data, as well as retrieval-augmented generation (RAG) that considers additional information from internal data sources. Another feature called “Continued pre-training” improves a model’s knowledge and reasoning abilities using a customer’s unlabeled data in their industry or field.
“We want you to be able to use Bedrock to complete actions like booking travel, filing insurance claims, deploying software… and this usually requires orchestration between multiple systems that operate in your organization,” Selipsky explained.
To further improve safety for AI models, Selipsky announced the introduction of Amazon Bedrock Guardrails, allowing customers to configure models to avoid certain topics, words, or responses based on responsible AI policies.
“We’re approaching the whole concept of generative AI in a fundamentally different way because we understand what it takes to reinvent how you are going to build with this technology,” Selipsky stated.
Overall, AWS continues to prioritize assisting organizations throughout the AI lifecycle, providing infrastructure, models, and applications. With the introduction of Amazon Q and the enhancements to Amazon Bedrock, the company aims to empower developers and businesses in their AI journey.
