How to Use Gemini AI

Gemini AI is a large language model from Google that can assist you with a gamut of tasks with its remarkable capabilities. Gemini AI is Google's most capable model yet, with their newest offering, Gemini Advanced, going neck and neck with competitors like GPT-4.

This powerful tool boasts a wide range of uses, from writing creative content to translating languages and answering questions in a comprehensive way. If you're new to Gemini AI, this guide will help you get started.

What is Gemini AI

Gemini AI is Google's latest and most advanced AI model, developed by Google DeepMind. Gemini is multimodal in nature, which means it can process and understand various types of data, including text, code, image, and potentially audio, and video.

Gemini AI isn't a single model, but rather a family of large language models from Google. These models are trained on massive datasets to generate human-quality text, translate between languages, answer your questions, and much more.

There are currently three model sizes of Gemini available: Ultra, Pro, and Nano.

Ultra is their most capable and largest model yet, which has shown exceptional performance on benchmarks, and is designed for complex tasks; it is available through a monthly subscription to Gemini Advanced that costs $19.99/month, which you can avail right now (with a 2-month free trial).

The Pro model is optimized for a wide range of tasks and is available widely through the Gemini chatbot (previously Bard), while the Nano model is the most efficient model, tailored for on-device applications, including smartphones. The Nano engine is the one that powers generative AI features on Pixel 8 Pro.

Gemini is also available for developers through API. For users who want to use Gemini as a chatbot like ChatGPT, Gemini (powered by Pro and Ultra) is the way to go. Gemini will also be available through a new Gemini app on Android and through the Google app on iOS.

Since the nomenclature is a bit confusing, here's a clear distinction again: Gemini, previously Google Bard, is the chatbot (powered by the Pro engine) that can be accessed for free, much like ChatGPT 3.5, and Gemini Advanced (powered by the Ultra engine) is the subscription-based model with advanced capabilities, like ChatGPT Plus that gives access to GPT-4.

You can find detailed comparisons between the two in our guide below.

Gemini (free) vs Gemini Advanced

Unsure if Gemini Advanced’s paid features are worth the upgrade? This post compares free Gemini and Gemini Advanced.

All Things HowShivam Malani

What Can You Do with Gemini AI?

Here are some of the things that Gemini AI can assist you with:

Content Creation: Gemini AI can write anything from poems and stories to marketing copy and website content. It can brainstorm ideas, suggest outlines, or write entirely from scratch.
Question Answering: Ask Gemini questions about pretty much anything and get informative, detailed responses. From history to science to pop culture, it's like having a super-smart encyclopedia at your fingertips.
Language Translation: Break down language barriers! Gemini can translate between multiple languages with impressive accuracy.
Code Generation: Gemini AI can even help with coding tasks. Provide descriptions of what you want your code to do, and it can help generate the corresponding code.
Generate images: It can generate images and infographics for you from just your descriptions.

How to Use Gemini AI

Whether you want to use Gemini for free or with a subscription, here's how you can do it.

To use the Gemini AI chatbot, navigate to the gemini.google.com web app from your web browser. If you go to bard.google.com (the previous address for accessing Google's chatbot), you'll be redirected to Gemini.
You should be logged in with your personal Google account or a workspace account for which access to the Gemini web app is enabled. To use Gemini Advanced, though, you need to have a personal Google account, as it's not available for Google Workspace accounts yet.

Chat with Gemini AI

If you're a Gemini Advanced subscriber, you can switch models from the top-left corner of the screen. Click the 'drop-down menu' and select your desired model. Whenever you visit the web app, the last model you used will be automatically selected.

Now, to chat with Gemini, go to the text box at the bottom of the screen and enter your prompt. You can enter a simple text prompt using your keyboard or dictate it using the microphone.

You can also upload an image to your prompt to chat with the AI about it by clicking the 'Upload Image' button.

Click the 'Submit' button to send your prompt.

Gemini will start responding to your query. You can stop its response before it has completed by clicking the 'Skip Response' button if you see that the response is not aligning with what you had in mind.

Edit the Prompt:

You can also edit the most prompt in the conversation with Gemini to get a refined response from the AI. To edit your prompt, go to the prompt and click the 'Edit Text' button that appears.

Then, edit your prompt and click the 'Update' button.

This will initiate a new response from Gemini.

View other Responses:

For the latest response from Gemini, you can also view other drafts that the AI crafted for your request. They might not be available every time, though. And they are definitively not available when using extensions with Gemini.

To see other responses, click on the 'Show drafts' button.

It will show other draft options after a couple of seconds. Click on the one you want to review.

To regenerate the responses, click the 'Regenerate drafts' button.

Modify, Double-Check, and Share the Response:

There are also additional options available with the response, like modifying, sharing, or double-checking it.

To modify the response, click the 'Modify Response' button and select how you want to modify it.

Click on the 'Share' button to share it with someone else, export it to your Docs, or open it in Gmail as a draft.

To double-check the response, click the 'Google' icon.

This will run a Google search for your query and evaluate Gemini's response based on the search results.
It will then highlight the content for which it found similar content during its search in green and provide a link for it. It's not necessary, though, that Gemini might've used the same link to produce its answer.

The content for which it found no similar content will be highlighted in orange.

The content that is not highlighted was not meant to convey factual information, or there's not enough information available to evaluate those statements.

Since it can be hard to trust an AI's answer, this is a great option that will let you evaluate which statements you can trust and which you should assess further.

Access your Chats:

To access your previous chats with Gemini, click the 'Expand menu' (hamburger icon) in the top-left corner of the screen.

The sidebar will expand. Here, you can access your previous chats that Gemini names automatically depending on the context of the chat.
You can also rename, pin, and delete a chat by clicking the three-dot menu icon that appears when you hover over it.

Click the 'New Chat' icon to start a new chat with Gemini, clear of the previous context.

Use Extensions:

You can also use extensions with Gemini AI if you're using a personal Google account; extensions aren't available with a Google Workspace account yet.

To access extensions, go to the menu on the left and click the 'Settings' button.

Then, select 'Extensions' from the flyout menu.

Here, you can enable/ disable the extensions you want to use.

Some extensions, like Google Flights, YouTube, Google Hotels, and Google Maps, will be enabled automatically.

To use the Google Workspace extension, you'll need to manually enable it and connect your Workspace account with Gemini, which will give it access to your docs, emails, drive, etc.

Generate Images with Gemini AI

Gemini AI can also generate images from your natural language prompts. All you have to do is describe the image you want to the AI.

Go to the prompt area and start your prompt with words like "Draw", "Generate", or "Create" and describe the image.
Click on the 'Submit' button.

The AI will generate a set of four images. Click on 'Generate More' for even more images.

You can download an image by hovering on it and clicking the 'Download' button.

You can also ask Gemini to generate images while you are prompting it to create other content, for example, a story, blog post, etc., by adding "and generate images for it" to your request.

Prompt Examples for Gemini

Gemini can help you with a variety of tasks, but your prompt will need to reflect what you're asking the AI for help with. You can send prompts like:

I'm planning a birthday party with a space theme. Give me suggestions for decorations, food, and activities.
What were the major causes of World War II? Explain it in a way that an elementary school student would understand.
Help me come up with a 30-day workout challenge I can do at home with minimal equipment.
I want to learn Spanish. What are some good resources or study methods you'd recommend?
Debug this error message: [error message]
Write an outline for a blog post [topic]
Write the opening scene of a mystery novel where a young detective investigates a disappearance at a carnival.

With Gemini Advanced, you can use more complicated prompts as the model is better at following complex instructions and retaining context from previous prompts.

I have a dataset with customer order information. Write Python code to identify the top 5 most frequently purchased items and determine their average profit margin.
I'm giving a presentation about sustainable business practices. : the latest research trends and provide 3 impactful case studies.
You are interviewing for a freelance project as a copywriter. Draft a cover letter template that includes a section for your portfolio, creative process, and rates.

When crafting prompts for Gemini, remember these points:

Gemini is multimodal. So, where required, pair your prompts with images or diagrams to make your interactions richer.
The first response might not be perfect. Treat it like a conversation and build on previous responses to get the desired outcome.
Explain the intended outcome (a blog post vs. an internal memo) so Gemini can adjust its style and format accordingly.
Provide Context. Some background or a small scenario adds richness to the prompt.
You can also experiment with the AI. Ask for different tones (formal, humorous, poetic, etc.) to shape Gemini's responses.
Be specific. Vague prompts yield vague answers. The more details, the better.
Practice open-ended prompts according to the situation. Sometimes you want facts (What's the capital of France?), but open-ended questions spark more creative responses.

Gemini AI is Google's most powerful AI model yet, capable of a wide range of tasks, from answering questions to generating images. Whether you use the free Gemini chatbot or subscribe to Gemini Advanced, you can use the chatbot to improve your productivity. With continued improvements and expanded capabilities, Gemini promises to be a useful AI assistant for both personal and professional use.