Oracle Primavera Unifier Version 26 Documentation
FAQs (General and AI-Specific)
Here are the answers to your most frequently asked questions.
AI and OCI
- Do Unifier and Oracle Cloud Infrastructure (OCI) Generative AI retain customer-provided data to fine-tune or train AI models?
- Do Unifier and OCI Generative AI retain customer-provided prompts and inputs used for inference on large language models (LLMs)?
- Do Unifier and OCI Generative AI share prompts and responses with third-party model providers such as Cohere, Meta, xAI, or Google Vertex AI?
No. Customer-provided data remains in the customer's Unifier database. OCI Generative AI does not use this data to train models. OCI Generative AI does not retain this data after the user interaction ends. The data is used only to generate a response to the customer's request.
No. OCI Generative AI does not retain customer input. When a user submits an inference request, OCI Generative AI sends the input to the LLM, the LLM generates a response, and OCI Generative AI returns the response to the user. OCI Generative AI does not store the input or the output. Customer-provided data and input are stored in the Unifier database for audit purposes.
No.
For more information about how OCI handles information, see the following:
- Guardrails for OCI Generative AI
- Data Handling in Generative AI
- Private Health Information (PHI) Identification Model
- Personal Identifiable Information
Notifications
Passwords
Proxies
Release Paths
New features are released twice a year (April and October) for Unifier cloud service and annually for Unifier on-premises. In addition, bug fixes are provided monthly for Unifier cloud service.
For more information, go to My Oracle Support (MOS) and see Doc ID KB858863.