Hands on with Google’s new Gemini app

Formerly known as Google Bard, Google’s AI chatbot boasts a new name and a host of new capabilities. More significantly, a new Gemini app can now replace Google Assistant on your Android phone, assuming your phone is set to US English. I won’t spend too much time diving into the confusing nuances of Google’s AI strategy (or seemingly lack thereof). Instead, I’ve left that scrambled chaos to more eloquent voices, and simply go hands-on with the Gemini app.

When dealing with an unruly topic, my first instinct is always to gather as much basic information as possible before testing features myself. In this case, I wanted to determine what benefits Google intends to deliver by replacing Google Assistant with the Gemini app for Android. In the face of ChatGPT and generative AI, Google is clearly burning rubber in an effort to launch its own AI offerings.

But what is the difference between Google Assistant and the Gemini app? I went right to the source and asked the Gemini app to enlighten me. You can see the app’s full response above, but in short, both are AI-powered tools with unique purposes. The Google Assistant we know and love manages basic tasks, controls smart devices, and performs as expected. Gemini is a large language model with extensive capabilities, including creative endeavors. It’s also still in its fledgling state, so you always have the option to swap back to Google Assistant as needed.

What began with voice assistants and online chatbots has officially consolidated into one powerful tool: an AI-driven chatbot and assistant. The app has some kinks to work out, but it’s a tool with immense potential. As of now, the Gemini app is only available for phones running Android 12 and above, and only in select regions, but the tool will be hitting iOS phones as well.

In practice, the Gemini app is a lot of fun. To initiate prompts, you can type, talk, or grab a photo and post a query. Anyone who has messed around with ChatGPT will find many of the same abilities as well as powerful Google integration.

You can ask Gemini to answer basic questions, craft a difficult message, or generate adorable images of unlikely animal interactions. I tested all of these (the last one quite excessively), and the app handled the prompts seamlessly. For the record, I did not forget my mom’s birthday; I’m not a monster.

The app can also help you interact with your immediate environment or provide insight into the things around you. Snap your phone’s camera at items ranging from toys to buildings and request Gemini identify the subject or offer guidance, like, “How can I use this?” or “Who designed this?” You can adjust your search as needed, offering more details to fine-tune your responses. Above, you can see how the app attempted to identify my desktop toy, and narrowed in its response as I offered insight.

I also took a photo of a mysterious jar of pickles from my fridge. The jar is mysterious because I can’t remember when or where I bought it, but I love the pickles, and I rationed them like toilet paper during the pandemic. I asked Gemini where I could buy more, and it turns out they’re readily available from many places, including the grocery store less than 2 miles from my home. Sure, I could have Googled this weeks ago, but why expend effort when an AI assistant will do it for you?

Fortunately, the Gemini app offers much more significant help as well. You can ask the app to synthesize information, which is done effectively. I submitted the link to an extensive article on the benefits of morning sunlight, and Gemini spit out a concise paragraph summarizing the important points. I tried the same approach on a 23-minute YouTube video, and the app kicked back another solid synopsis.

You can even ask Gemini to evaluate Taylor Swift’s entire discography and determine the saddest song she’s written (in case you have a bottle of wine to kill). Oddly, this query produced a response partially written in a different language. I asked Gemini why this was the case and the app apologized, citing that it’s still under development.

If you do get a response in a different language, well, you can also ask Gemini for a translation of foreign text. I’ve recently lamented to Android Authority‘s own Rita El Khoury about the remorse I feel knowing only 1.5 languages in light of her multilingual abilities. I am happy to add assistance to my pocket.

In my experience, the AI capabilities of Gemini are right on par with those of ChatGPT 4. It’s comprehensive, responsive, and friendly (almost to an unsettling degree). It’s very easy to imagine how this app can be used regularly.

But what about the dethroning of Google Assistant? Once you install Gemini, the app replaces Google Assistant with a brand-new interface, and though it brings some powerful new potential, it also makes some unfortunate cuts.

Cutting to the chase, the Gemini app is not yet capable of replacing Google Assistant, at least not for anyone who uses the latter religiously. If your hectic schedule has you chiming, “Hey Google, add x event to my calendar” on a regular basis, Gemini isn’t for you. According to the company, Google will continue to add support.

What Gemini can’t do that Google Assistant can:

• Add calendar events
• Play music
• Assist with reminders
• Accurately report live sports scores

Like ChatGPT, Gemini will sometimes provide responses that are not correct. I asked the Gemini app for the current leader in a live golf tournament and received a response that included a real player and potential score. However, the actual score was incorrect.

The app does offer the option to verify your query responses by tapping the colorful Google G. Gemini will then highlight the text to identify whether a Google search produces the same content.

• Set timers
• Provide the weather
• Integrate with smart home devices (lights, TV)
• Provide sports schedules
• Answer basic questions
• Make phone calls

Notably, the Gemini app does not offer results instantaneously, presumably because it sifts through much more information to provide broader results. For some requests, like “create a story about the Notre Dame Football team with the same moral as the three little pigs,” I expect to have to twiddle my thumbs. But when simply asking for the current time in New York City, I appreciate Assisstants’s expediency. As entitled as it feels to complain about, it’s a difference worth noting when comparing Google Assistant and the Gemini app.

Gemini also degrades the hands-free experience of Google Assistant. Beyond your first prompt, you need to repeat “Hey Google” or physically press send on follow-up questions or requests. On Google Assistant, you can enable continued conversations, but on Gemini, that feature is not yet available.

As mentioned earlier, installing the Gemini app replaces Google Assistant on your Android phone. That means “Hey Google” initiates the new app, as does long-pressing the power button or swiping in from the corner of your screen. You are anticlimactically consciously uncoupled from the former assistant. If you give Gemini a go and miss your old companion, you can roll back the change with no sweat. However, you can’t use both simultaneously. The Gemini app does not work as a standalone tool and resets if you return to Google Assistant.

This is simply obnoxious. It’s clear that Google intends to send its old assistant out to pasture, but denying users access to both tools makes no sense until that day comes. Given that Gemini cannot yet complete some of Google Assistant’s most commonly used tasks, it would be very beneficial to have access to both options.

Without getting too riled up, this oversight isn’t just silly; it’s classic Google. The Gemini app is a powerful tool with exciting potential, but its current rollout is nonsensical.