The Google app for iOS is set to introduce an innovative artificial intelligence (AI) feature designed to assist users in deciphering complex text found in articles and web pages. Starting Tuesday, this new capability, known as Simplify, began rolling out to users. Utilizing Gemini AI models, this feature enables individuals to select specific text for an AI-generated rewrite that clarifies intricate language, jargon, and specialized technical terminology. Simplify has been created by Google Research, and the company has released a technical paper that outlines the development process.
New Text Simplification Feature for Google App on iPhone
In a recent blog post, Google provided insights into this feature now available to iOS users. The rollout is being conducted in phases, which means it may take several days for all users of the Google app on iOS to gain access to the AI functionality.
The AI tool is particularly beneficial for users exploring unfamiliar topics and facing challenges due to the intricate terminology involved. Once they have access, users can highlight any specific portion of text, prompting the AI to rewrite it in a more accessible language.
After highlighting the text, users will see the Simplify option located at the bottom of the screen within the “More actions” panel, appearing as the first option on the right. The feature is identifiable by an icon depicting the letter ‘A’ encircled by two curved arrows. Selecting this icon activates the text simplification process.
A key advantage of Simplify is that it generates the reformulated text directly on the same page, allowing users to remain in their current article or research document while gaining clarity on complex phrases. The tool operates on the Gemini 1.5 Pro AI model, developed by Google Research.
In a detailed technical paper, researchers from Google emphasized that fidelity assessment was central to the AI tool’s design. They focused on ensuring that the essence of the original text is preserved, addressing potential errors like information loss and inadvertent insertion of unrelated words, as well as ensuring accuracy in meaning before the final text is generated.
