Google adds live search to Gemini

In response to OpenAI’s advancements, Google has launched a feature aimed at enhancing its Gemini model by integrating real-time data from Google Search. This initiative is designed to expedite query processing and reduce instances of inaccurate information, often referred to as “false positives.”

As reported by TechCrunch, following OpenAI’s introduction of a live web search capability for ChatGPT, Google swiftly introduced the “Grounding with Google Search” feature for Gemini. This new functionality enables Gemini to access current, relevant information from Google Search, thereby improving the accuracy and reliability of its responses. This is particularly beneficial for complex queries that require up-to-date data.

The “Grounding” feature allows Gemini to reference real-time information, beyond the static data typically stored in its database. When users pose questions, Gemini automatically retrieves pertinent information from Google Search, providing more credible answers and citing direct sources for user verification. This approach significantly mitigates the risk of “hallucination,” where an AI might generate misleading or incorrect information, and enhances transparency by allowing users to trace the source of the information.

Currently, this feature is not available in Europe, the UK, and Switzerland; however, Google has indicated plans to roll it out in these regions shortly. To access the new feature, users must enable the “Show suggestions from Google Search” option in their settings. Once activated, search results will appear beneath Gemini’s responses, facilitating direct access to information sources.

It is important to note that utilizing the “Grounding with Google Search” feature may lead to increased operational costs and potential delays in response times. To address these concerns, Google provides developers with a Dynamic Retrieval option. This permits developers to regulate how Grounding is utilized by establishing a necessary threshold to optimize costs. Developers can adjust this threshold based on prediction scores, where higher values correlate with stronger levels of Grounding.

The competitive landscape between Google and OpenAI continues to intensify, as both companies strive to refine their AI technologies to emerge as the foremost reliable and up-to-date search engines for users.

Related posts

Google launches Gemini 2.0 – comprehensive AI that can replace humans

NVIDIA RTX 5090 can be 70% more powerful than RTX 4090?

iOS 18.2 launched with a series of groundbreaking AI features