In the competitive landscape of AI chatbots, Google aims to keep users engaged by providing exclusive content, tailored responses that reflect individual online habits.
On Thursday, the tech giant unveiled Gemini with personalization, an innovative “experimental feature” for its Gemini chatbot applications that allows Gemini to leverage other Google apps and services for customized responses. According to Dave Citron, product director for Gemini, this feature can utilize users’ activities and preferences across Google’s ecosystem to offer personalized answers to inquiries.
“These enhancements are crafted to make Gemini feel less like just a tool and more like a seamless extension of yourself, poised to anticipate your needs through genuine personalized assistance,” Citron shared in a blog post to TechCrunch. “Initial users have found Gemini with personalization beneficial for brainstorming and receiving tailored suggestions.”
The Gemini with personalization feature will initially integrate with Google Search, with plans to expand to additional services such as Google Photos and YouTube in the upcoming months. This move comes as chatbot developers, including OpenAI, strive to distinguish their assistants by offering distinctive features. OpenAI has recently introduced the capability for ChatGPT on macOS to edit code directly within supported applications, while Amazon is preparing to unveil an enhanced “agentic” version of Alexa.
Citron stated that Gemini with personalization operates on Google’s experimental Gemini 2.0 Flash Thinking Experimental AI model, which is a so-called “reasoning” model that can assess whether personal data from Google services, like a user’s Search history, can improve the relevance of answers. Specific inquiries informed by preferences, such as “Where should I plan to vacation this summer?” and “What new hobby would you recommend?”, will benefit the most, Citron explained.
“For instance, you can request restaurant suggestions, and it will consider your recent food-related searches,” he noted, “or seek travel advice, and Gemini will reply based on previously explored destinations.”
However, this raises privacy concerns as there’s potential for Gemini to unintentionally reveal sensitive information.
To address these concerns, Google is making Gemini with personalization an opt-in feature and restricting it for users under 18. Citron indicated that users will need to grant permission before the Gemini can access their Google Search history and other apps, providing transparency regarding which data sources are utilized for customizing responses.
“When you’re utilizing the personalization feature, Gemini will show a clear notification that includes a link to easily disconnect your Search history,” Citron mentioned. “Gemini will only have access to your Search history once you opt into Gemini with personalization, following your approval to connect to your Search history, and when your Web & App Activity setting is enabled.”
Gemini with personalization will be available to users on the web (excluding Google Workspace and Google for Education users) starting Thursday and will gradually roll out to mobile platforms thereafter. The feature will support over 40 languages across “the majority” of countries, according to Citron, with certain regions, such as the European Economic Area, Switzerland, and the U.K., excluded from the launch.
Citron hinted that the feature might not remain free indefinitely.
“There may be future usage limits,” he stated in the blog entry. “We’re committed to gathering user feedback on how to best leverage this capability.”
New Models, Connectors, and Enhancements
To encourage continued use of Gemini, Google has introduced enhanced models, research functionalities, and app integrations for the platform.
Subscribers to Gemini Advanced, Google’s $20-per-month premium service, can now access a standalone version of 2.0 Flash Thinking Experimental that supports file attachments; integrations with apps such as Google Calendar, Notes, and Tasks; along with a context window of 1 million tokens. This “context window” refers to the amount of text the model can process at one time—1 million tokens equals approximately 750,000 words.
Google explained that this latest iteration of 2.0 Flash Thinking Experimental is not only faster but also more efficient than its predecessor, with improved capabilities for handling requests that involve multiple applications, such as “Find an easy cookie recipe on YouTube, add the ingredients to my shopping list, and locate grocery stores still open nearby.”
In what seems like a response to OpenAI’s recently launched tools for in-depth research, Google is also enhancing Deep Research, a Gemini feature that compiles web-based reports on various topics. Deep Research now makes its “thinking” process transparent and employs 2.0 Flash Thinking Experimental as its primary model, which is expected to produce higher-quality, more detailed, and insightful reports, according to Google.
Deep Research is now available for free trial to all Gemini users, with increased usage limits for customers of Gemini Advanced.
Moreover, free Gemini users will have access to Gems, Google’s customizable topic-focused chatbots that previously required a subscription to Gemini Advanced. With upcoming updates, all Gemini users will also be able to interact with Google Photos, enabling features like retrieving photos from recent trips, according to Google.
Compiled by Techarena.au.
Fanpage: TechArena.au
Watch more about AI – Artificial Intelligence
