Google LLC today announced the rollout of “grounding” its Gemini artificial intelligence models using Google Search, which will allow developers to get more accurate and up-to-date answers to using search results.

New updates are available for AI Studio and the Gemini API. They enabling the model to provide not only more accurate answers, but also feedback from sources in the form of online support links and suggestions that direct users to search results in the answer.

By earthing Using the search results model, developers can import new search results data to correct issues that might arise from outdated training data. When a user asks an AI a question conversationally, it responds with its best knowledge, but the training can only respond with the information it has up to a certain point.

The response could contain elements that are several months old and therefore less precise and less desirable. When a user makes a query with grounding enabled, Google said, the service uses its search engine to find up-to-date and complete information. relevant to the query, then sends it to the model.

Use sground research The AI ​​responses will also allow the model to respond with more context and generate better results. As a result, the information will be richer and filled with better resources than the model alone. With the addition of links to potential research resources, users will also have the ability to perform additional research.

Google Search-backed data can also help ensure that AI applications provide users with more factual information. According to the company, this can also reduce hallucinationswhen an AI model’s response contains false or misleading information presented with confidence.

Using the Application Programming Interface, developers define a configuration that will determine when grounding is more or less likely to be beneficial for a query. In this setting, each prompt is assigned a value between 0 and 1 that predicts whether a prompt will benefit from grounding and the developer can set a threshold to enable grounding. A higher resulting number means the ground will be more useful. The default threshold is 0.7. Google suggested developers experiment to see which setting best suits their app.

Google said this configuration exists because Grounding with Search would result in slower responses from Gemini models. Users may notice this slowdown in apps and it could make their experience worse. However, for prompts that would benefit, the slower response would likely be negligible.

There is also the question of additional costs associated with using the new update. Grounding is available to test for free in Google AI Studio. In the API, developers can access the tool with the paid tier for $35 for 1,000 queries founded.

Image: SiliconANGLE/Bard Image Generator