Before Siri integrates with ChatGPT, Google has entered Apple's territory: iPhone users can now use the Gemini App

Author: Li Dan

Source: Hard AI

Just as Apple's Siri was preparing to integrate ChatGPT, Google quietly entered Apple's App Store, and iPhone users can now use Google's AI assistant, Gemini.

On Thursday, November 14th, Eastern Time, the iOS version of Google's AI chatbot, the Gemini App, was launched on Apple's App Store, currently ranking 37th in the productivity app category. This free app displays only a chat window and a list of past chats. iPhone users can ask Gemini questions via text, voice, or camera to receive answers.

The iOS version of the Gemini App supports prompts in 35 languages, and users can converse with Gemini in 12 languages through the Gemini Live feature in the app, including English, Spanish, French, German, Hindi, Portuguese, Arabic, Italian, Indonesian, Japanese, Turkish, and Vietnamese. Google plans to soon support more languages.

The iOS version of the Gemini App shares some features with the Gemini App on the Google App Store and the version accessible on the Gemini website, such as users being able to request recipe recommendations or seek learning assistance, but Gemini Live is a new feature exclusive to the iOS app. It makes the chat with Gemini more interactive and conversational, similar to the voice mode of ChatGPT.

A few weeks ago, Google launched Gemini Live for Android devices, marking the first time this feature is available for iPhone users. David Pierce, a contributing editor at The Verge, stated that he conducted some brief tests, showing that Gemini Live runs very well on the iPhone, displaying simultaneously on the iPhone's Dynamic Island and lock screen when using this feature.

Pierce believes that the significance of the Gemini App lies in its icon being placed on the user's home screen, becoming a point that users can assign to an action button or quick access on their phone. With just a tap, in half a second, users can chat with the Gemini bot. This kind of access and the muscle memory it helps build are crucial for any company looking to encourage users to develop a habit of chatting with a bot.

Of course, like all other non-Siri chatbots, Gemini has some significant limitations on the iPhone; users cannot change settings or access other apps while using it. However, it can access other Google apps, which remains a significant advantage for Gemini For example, users can ask Gemini about information they might have left in other Google accounts like Gmail, Maps, YouTube, and Calendar. When users request Gemini to play music, it can launch YouTube Music, and when asking for directions, it can show users Google Maps. This is also part of what Google hopes to achieve with the Android system and what Apple aims to accomplish through Siri: using AI to make all content on the phone more interactive and accessible. Of course, if users do not use AI robots, this goal cannot be achieved, so both Apple and Google will compete for users' home screens.

According to the introduction of the Gemini App on iOS, users can also generate images using Google's text-to-image model Imagen 3, suggesting it can be used to generate previously controversial character images. Since iPhone users can try Google's most powerful AI model, Gemini Advanced, for free for a month, if they have already subscribed to Gemini Advanced, they can directly access the model within the App.

In February this year, Google's Gemini launched a text-to-image tool that was criticized for generating character images that did not match reality. Netizens complained on the social media platform X, formerly known as Twitter, that Gemini blatantly "distorted history," with requests to generate images of American founding fathers and medieval English kings resulting in depictions of Black individuals. Some requests for images of German soldiers from World War II in 1943 led to Gemini generating soldiers who were Asian and Black.

Google had to decide to take down the character image generation feature, stating that part of the issue stemmed from the model becoming "much more cautious than we expected" over time, misinterpreting some harmless prompts as sensitive words. It wasn't until the end of August that Google began allowing users of the three versions of Gemini Advanced, Business, and Enterprise to re-enable the character image creation feature.

It is worth mentioning that last week, Apple launched the iOS 18.2 public beta, in which Siri integrated ChatGPT, meaning all iPhone users will be able to use ChatGPT-enhanced Siri in the coming weeks. A new round of competition between Gemini and ChatGPT is about to unfold on the iPhone