Here is how Google’s new AI tool, Gemini works

2023 was regarded as the year of AI. Now it is 2024 and the ‘AI craze’ seems to not be dying down soon. Amongst other tech giants, Google has now rolled out its own AI chatbot.

Launched last year as Bard, Google announced last week that its chatbox will now be called Gemini, the name of the Large Language Model (LLM) powering the AI chatbot. According to a statement shared with Technext, the name change is meant to reflect the “advanced tech at its core”. 

Read here:

From Bard to Gemini: Google renames chatbot to reflect improved capacity

Building on the rebrand, Google rolled out new Gemini and Gemini Advanced apps for Android and iOS users. With this, we have downloaded and used the iOS version of Google Gemini. Here is how it works.

Review of Google Gemini 

Simply put, Google Gemini is a family of language models. To access the model on IOS, search ‘Google Gemini’ on the Apple Store and download the app with the icon in this image.

After downloading, you will be able to access the application’s free functionality with or without signups. For Nigerian users, the Google Gemini pro version goes for N5,900.00/week and N14,900.00/month.

One would have to commend Google for Gemini’s variety of features and how they are sectioned on the app. Apart from the chat section, where you can ask the AI chatbot general questions and get responses, it has a section of ‘expert.’ Here, there are varieties of personalities in different fields, to choose from.

For instance, if you have any tasks/questions pertaining to food and cooking, there are options of Sushi Chef, Japanese Cuisine, Ramen Expert, and others to assign questions/tasks. These features aren’t obtainable in the free version of other AI chatbots.

I asked the Sushi Chef about the procedure of making shawarma. After giving an initial caveat noting that its field of expertise is sushi and not shawarma, it gave me a wholesome and valuable insight into how shawarma can be made, which was impressive. 

Other ‘experts’ fields include Religion, Career, Health, Fitness, Anime and others. Interestingly, there’s a section modelled after Elon Musk’s character where you get to explore space, technologies and futuristic visions.

Google Gemini AI image generator 

The AI image generator is for Pro users, but I was able to access it as a new user for a single time. With proper descriptions, images can be generated through this feature called ‘Generate Image’ on Google Gemini.

As someone that does not fancy stuff like this, I instructed the image generator to create an image, using this description:

Positive Prompt: A young struggling man in Lagos, Nigeria

Negative Prompt: A cartoonish image

And I got the below image:

That’s an interesting response, despite an unelaborate description 

Google Gemini Tools

The Google Gemini AI chatbot has a more sophisticated feature called Tools. Under tools, you can find YouTube Summary, Ask YouTube, Document Summary, and Image Recognition. All these features are however available only for Pro version subscribers. 

For the YouTube Summary feature, a user can paste a YouTube video link, and choose a language they want to summarize it in and boom, it will generate a summary of the video. Also, the Ask YouTube tool answers all questions regarding YouTube videos in the same format.

The Document Summary gives a concise overview of PDF files while the Image Recognition tool looks out to features in images and describes them.

I tested the Image Recognition with this image and I got the below response. 

Response from Image Recognition:

“This is an image of a man’s head and upper shoulders encapsulated within a circular frame. The man has his eyes closed and his head slightly tilted backward, with his lips parted as if he is in the middle of speaking, singing, or possibly exhaling. He has a full beard and it appears he’s indoors with a nondescript background.”

Related post:

5 AI tools to help you get the best auto-generated AI images for your content

I pressed further by asking who the particular person in the image was, and the tool retorted:

“I’m unable to see the image you’re referring to as I don’t have the capability to process or interpret images. If you could provide a description of the image, I’d be happy to help based on that description!”

Google Gemini History 

The last and probably one of the most interesting features on the app is History. Here, you have access to all your activities on the application. Also, History has sections for different features like chats, images and tools. Which makes the whole app interesting for documentation.

Other perspectives 

At least at these initial stages, Google seems to be serious and intentional about rivalling or taking over the AI market from ChatGPT. The free mode features, looking at the expert/field parts are expected to garner more AI users/enthusiasts to Gemini’s side than its competitor. 

However, Google Gemini seems to lag with responses. Compared to its competitor, it does not seem to enjoy swift responses to questions and assignments. 

With the already deployed simple and easy-to-use interface, if Gemini could improve on fast response, the AI tool could be unstoppable.