Google touts ‘enterprise-ready’ AI with more facts, less pretense

Vertex AI, a Google Cloud development platform that enables companies to build services using Google’s machine learning and vast language models gaining new opportunities to prevent apps and services from transmitting wrong information. Once Vertex AI solutions are generally available Grounding with Google Search in May — which allows models to pull information live from the internet — Google has now announced that customers will also have the ability to enhance its services’ AI output with specialized third-party data sets.

Google says the service will leverage data from providers like Moody’s, MSCI, Thomson Reuters and ZoomInfo, and that grounding in third-party data sets will be available in the “third quarter of this year.” It’s one of several modern features Google is developing to encourage organizations to adopt its Generative, enterprise-ready AI solutions. reducing the frequency with which models provide misleading or wrong information.

Another is “high-fidelity mode,” which allows organizations to source information for generated results from their own corporate data sets rather than from Gemini’s broader knowledge bank. High-fidelity mode is supported by the specialized version of Gemini 1.5 Flash and is now available in preview via Vertex AI Experiment Tool.

Vector search, which lets users find images based on similar graphics, is also being expanded to include hybrid search support. The update is in public preview and lets you combine vector searches with text searches using keywords to improve accuracy. The connection to Google Search will also soon provide a “dynamic fetch” feature that automatically chooses whether information should come from established Gemini datasets or Google Search for suggestions that may require frequently updated resources.