Oracle Integrates Generative AI Throughout Technology Stack for Widespread Enterprise AI Implementation

Oracle has unveiled its Oracle Cloud Infrastructure (OCI) Generative AI service, marking a significant step towards facilitating widespread enterprise adoption of AI at scale. The fully managed service seamlessly incorporates large language models (LLMs) from Cohere and Meta Llama 2, catering to diverse business use cases. The OCI Generative AI service now boasts multilingual capabilities supporting over 100 languages, an enhanced GPU cluster management experience, and flexible fine-tuning options. This service is accessible both on the Oracle Cloud and on-premises via OCI Dedicated Region.

Greg Pavlik, Oracle Cloud Infrastructure’s Senior Vice President for AI and Data Management emphasized Oracle’s commitment to solving real-world business challenges by integrating generative AI throughout the entire technology stack. Rather than providing a toolkit for customers to assemble, Oracle offers a suite of pre-built generative AI services and features that work cohesively to help customers address business problems efficiently.

To simplify customization of generative AI models for tasks such as text generation and summarization, Oracle is making the latest models from Cohere and Meta Llama 2 available in a managed service accessible via API calls. Customers can securely embed generative AI into their technology stack, ensuring tight data security and governance. The models can also be refined using retrieval augmented generation (RAG) techniques, allowing customers to tailor them to their unique internal operations.

In beta, the OCI Generative AI Agents service with an RAG agent combines large language models with enterprise search built on OCI OpenSearch. This enables users to interact with diverse enterprise data sources through natural language without specialized skills. The beta release supports OCI OpenSearch, with upcoming releases supporting a broader range of data search and aggregation tools, including access to Oracle Database 23c and MySQL HeatWave.

Oracle’s approach to generative AI involves embedding it across every layer of its technology stack, from applications and converged databases to infrastructure. This integration spans its cloud applications portfolio, including ERP, HCM, SCM, and CX. Oracle aims to empower customers to leverage generative AI innovations within their existing business processes. Generative AI capabilities are also being embedded into Oracle’s database portfolio through features like Autonomous Database Select AI, allowing customers to build their own AI-powered applications.

To support customers in building, training, deploying, and managing LLMs, Oracle is expanding the capabilities of OCI Data Science. The upcoming OCI Data Science AI Quick Actions feature, in beta next month, enables no-code access to various open-source LLMs, including providers like Meta and Mistral AI.

In summary, Oracle’s strategy involves integrating generative AI comprehensively across its offerings, creating a powerful combination of AI infrastructure and cloud applications to simplify the deployment of generative AI in customer workflows.