Google’s AI has gone through a couple of stages thus far. First, Google launched “Bard” as a direct competitor to ChatGPT. Soon after, the company rebranded to “Gemini,” encompassing everything from its chatbot to its LLMs. During The Android Show: I/O Edition, Google revealed its newest AI iteration: Gemini Intelligence. It might sound suspiciously similar to another company’s AI suite, but Gemini Intelligence offers a number of unique features—at least, according to Google.
Gemini Intelligence’s new agentic abilities
Google is positioning “Gemini Intelligence” as an agentic assistant. While the company already offers agentic capabilities, like ordering cabs through Uber or takeout through DoorDash, the new experience should have more abilities. For example, Google says Gemini can book you a front-row bike in your spin class, or order the books you need for a class after finding the syllabus in your Gmail inbox.
Google is also touting Gemini’s image context for automation. In one example, you could pull up your grocery list in your notes app of choice, then ask Gemini to add the items to your shopping cart. In another situation, you could take a photo of a travel brochure you find in your hotel’s lobby and ask Gemini to find you a tour for your group that matches the experience in the literature.
Gemini Intelligence offers more advanced autofill
Autofill is one of those features I definitely take for granted. It’s not always perfect (please don’t enter my phone number in the credit card form), but when it works, it saves you a ton of time filling out digital paperwork. According to Google, Gemini Intelligence is upping the ante with autofill, with the goal of filling in just about any information in any form.
The big example here concerns traveling: Imagine you’re buying a plane ticket, and you need to fill out your identifying information. While traditional autofill can help with your name, phone number, email, and so on, you usually need to find your passport (or a picture of your passport) to fill out that section. Google says Gemini Intelligence can tap into Personal Intelligence to autofill secure details like your passport information. You’ll see a button appear marked “Passport,” which, when tapped, enters in all of those details.
Google says the feature is “strictly opt-in,” so you have the choice whether or not to connect to Gemini Intelligence in these instances. You also have the option to disable it at any time.
“Rambler” improves dictation on Android
Gemini Intelligence is also aiming to improve dictation, specifically by polishing up our often unpolished thoughts. To do so, Google is introducing a new feature called “Rambler”: When you dictate to Gemini Intelligence, instead of writing down all of your “ums” and “uhs,” Rambler tries to jot down only what you intend to say. That includes when telling the AI you made a mistake. If you say something like “On my grocery list, I need three bananas, one orange juice, a gallon of milk, (oh wait, never mind, I have milk) and a loaf of bread,” Rambler should only write down “On my grocery list, I need three bananas, one orange juice, and a loaf of bread.” You can also ask Rambler to adjust the formatting of the dictation, so you can turn that string of grocery items into a bulleted list with emojis.
What do you think so far?
According to Google, there’s a visual difference between Rambler and standard dictation, so you should always know when the feature is on. The company says the audio is only used to transcribe it in real time, and isn’t stored after the fact. Google also says Rambler supports multiple languages, so you can use dictation while switching languages without having to start and stop dictation.
This is the feature that piqued my interest the most: Gemini Intelligence will, supposedly, offer users the chance to build their own widgets for the home screen. Rather than wait on app developers to make widgets that may or may not do what you what them to, you can ask Gemini Intelligence to build those widgets on your behalf.
Google says “Create My Widget” lets you ask Gemini to make widgets from prompts like “show me upcoming concerts at Madison Square Garden,” “display wind speed and rain for Golden, CO,” or “suggest new meal prep recipes at the start of each week.” Should the feature work as advertised, you can generate custom widgets that are uniquely tailored to your needs and interests.
