Week 4: Google Gemini

Learning objectives

By the end of this week, you will be able to:

1. Navigate Gemini's web and mobile apps with a screen reader, using the same exploration skills built in earlier lessons.
2. Explain what Gemini is, who builds it, and how it fits the large language model vocabulary introduced earlier in this course.
3. Explain what it means for an AI tool to be integrated with other apps, using Gemini's connection to Gmail, Google Docs, and Google Drive as the example.
4. Describe how a learner could use Gemini to help with a task inside Gmail, Docs, or Drive.
5. Describe a general, reliable screen-reader workflow for starting a conversation in Gemini, reading its response, and locating integration-related controls.
6. Compare Gemini to Claude and ChatGPT on at least three dimensions, including conversation quality, app integration, and mobile experience.
7. Explain why this course avoids naming exact, current menu paths or keystrokes for Gemini's integration features, and describe a reliable alternative.
8. Choose, with reasons, which of the three tools covered so far might fit a specific real task best.

Lesson

What is Google Gemini?

Google Gemini is a large language model chatbot built by Google, the same company behind the Google Search engine, Gmail, Google Docs, and the Android operating system. Like Claude and ChatGPT, which you explored earlier in this course, Gemini is a conversational AI assistant. You type or speak a message, called a prompt, and Gemini generates a response using patterns it learned from enormous amounts of text during training. Earlier lessons introduced the term large language model, or LLM, for this kind of AI system, along with related vocabulary such as parameters, tokens, context window, and hallucination. Every one of those terms applies to Gemini exactly as it applied to Claude and ChatGPT, because all three are built on the same general LLM technology, even though each company trains and fine-tunes its own model differently.

That shared foundation also means Gemini's honest weaknesses are the same general weaknesses earlier lessons taught you to watch for in any AI tool. Gemini can still hallucinate, confidently stating something false or invented. It can still reflect bias picked up from its training data. It still has a context window, a limit on how much of a conversation it can actively consider at once. None of this makes Gemini worse than Claude or ChatGPT; it means the verification habits you already built, reading carefully with your screen reader and cross-checking important claims against a separate, reliable source, apply here without any changes.

The most important thing to understand about Gemini going into this week is which company built it and why that matters. Google has spent more than two decades building products that already sit at the center of many people's digital lives: its search engine, its email service Gmail, its word processor Google Docs, its spreadsheet program Google Sheets, and its file storage service Google Drive. When Google built Gemini, it did not build a stand-alone chatbot with no other connections. Instead, it built Gemini to work alongside the products it already owns. That single design choice is the main thing that sets Gemini apart from Claude and ChatGPT, and it is the reason this week's lesson spends real time on integration, a term defined in detail later in this lesson.

Gemini exists as a website you can open in a browser, as a mobile app for phones and tablets, and, as you will read shortly, as an assistant built directly into some Android phones. Like Claude and ChatGPT, Gemini has different versions or models you can sometimes choose between, usually described as faster, lighter options for quick answers and larger, more capable options for harder tasks. You do not need to memorize specific model names for this course, since Google updates them over time, the same caution that applied to Claude's and ChatGPT's model names earlier in this course. What matters is understanding the general pattern: a faster option trades some capability for speed, and a larger, more capable option trades some speed for depth.

Navigating Gemini's web app with a screen reader

Opening Gemini for the first time with your screen reader should feel familiar, because the underlying exploration skills are exactly the ones you have been building throughout this course and used directly with earlier tools. A Gemini page, like most modern web apps, is built from standard web page structures: headings, landmarks, buttons, and edit fields. Your screen reader does not need Gemini-specific training to read these structures; it needs you to use the same general commands you already know. Before doing anything else on a new or updated Gemini page, it is worth opening your screen reader's list of links or interactive elements, sometimes called an elements list, which lets you look at every heading, link, button, and form field on the page in one organized list, sorted by type. Most screen readers offer some version of this feature; check your own screen reader's shortcut list for the exact key if you are not sure. This single feature is your most reliable starting point whenever an interface looks unfamiliar, and that includes Gemini.

From your screen reader's list of elements, or by browsing the page directly, you can use your screen reader's heading navigation to get a sense of how the page is organized, its button navigation to find controls such as a button to start a new conversation or open a menu, and its edit-field or form-field navigation to find the message box where you type your prompt. This is the same pattern you practiced with earlier tools in this course. Because Gemini, Claude, and ChatGPT are all built as modern web applications, they tend to share a similar general shape: a message box near the bottom of the page or screen, a send control near that box, and the conversation history displayed above it, readable with Say All, or your screen reader's continuous reading command, once you land inside it.

Landmark navigation is worth building into this same habit. Your screen reader's landmark or region navigation moves from landmark to landmark, such as a main content region, a navigation region, or a form region, and can help you jump past repeated header or menu content straight to the part of the page where the actual conversation lives. Combining landmark navigation with heading navigation and your screen reader's list of elements gives you three independent ways to orient yourself on any Gemini page, so if one approach does not immediately locate what you need, you have two more reliable fallbacks rather than guessing or arrowing through the whole page line by line.

One general pattern worth knowing before you explore Gemini yourself: many chat-style AI apps place the message box as the last, or one of the last, edit fields on the page, since it usually sits at the bottom of the visible layout. Repeating your edit-field navigation command from the top of the page, or jumping toward the bottom of the page first with your browser's end-of-page command before using edit-field navigation, can sometimes get you to the message box faster than starting from the very top. This is a general habit, not a guaranteed shortcut, since exact placement can differ or change over time. If a specific feature described in this lesson ever stops matching what you hear, treat that as completely normal for fast-changing web apps, and fall back on your screen reader's list of elements and its heading, button, and edit-field navigation to relocate what you need. That fallback skill matters more than any single memorized shortcut.

Once you have located the message box, type your question or request as you would with Claude or ChatGPT, then find the send control, most often reachable with your screen reader's button navigation until you hear something like "Send" announced, or simply by pressing Enter from inside the message box, which submits the message in most chat interfaces you have used so far in this course. After sending, Gemini's response typically appears above or after the message box. Move to it using heading navigation, landmark navigation, or by continuing to Tab or arrow through the page, and once you reach the start of the response, use Say All, or your screen reader's continuous reading command, to hear it read from that point straight through.

Gemini on mobile

Like Claude and ChatGPT, Gemini has a dedicated mobile app you can download for iPhone or Android, and the general exploration skills you built for those two tools carry over directly: on iPhone, VoiceOver's rotor and swipe gestures let you move by headings, buttons, and text fields; on Android, TalkBack offers comparable gestures and a similar reading-by-element model. What is different about Gemini on mobile, compared to Claude and ChatGPT, is that Gemini is not only available as its own separate app. On many Android phones, Gemini is built directly into the phone itself, offered as the phone's built-in AI assistant, similar in role to how Siri functions on an iPhone.

This distinction matters because it changes how you might reach Gemini day to day. Instead of always opening a dedicated Gemini app icon, you may be able to reach Gemini through an assistant gesture or button built into the Android phone itself, such as a long-press of a hardware button or a voice trigger phrase, depending on the phone, its Android version, and its settings, all of which can vary by device and change with software updates. Because these specifics vary so widely across different Android phones and change over time, this course cannot give you one universal, guaranteed keystroke for reaching Gemini as a phone assistant. Instead, the reliable approach is the same exploration habit from this whole course: use TalkBack's own exploration gestures to look through your phone's settings and installed apps, or ask Gemini itself, once you find it, how to reach it more quickly next time.

It is worth checking your phone's own accessibility settings menu as part of this exploration, since many Android phones list their built-in assistant, along with how it is triggered, inside a settings section related to the assistant or to accessibility features generally. TalkBack can read through settings menus the same way it reads any other screen, heading by heading and item by item, so this exploration does not require sighted help; it simply requires applying your existing TalkBack navigation skills to a new part of the phone.

Once you are inside the Gemini mobile app or the Gemini assistant experience on a phone, the reading experience should feel familiar from your work with Claude's and ChatGPT's mobile apps earlier in this course: a message field, a way to submit your message, and a scrolling conversation history you can read with your screen reader's standard reading commands, continuous reading on both iPhone and Android included. Mobile AI apps also often support speaking your prompt instead of typing it, using the phone's own dictation feature, a genuine accessibility advantage of voice input highlighted earlier in this course with another tool, and the same advantage applies equally well to Gemini.

Gemini's integration with Gmail, Google Docs, and Google Drive

Integration is this week's most important new term, and it is worth defining precisely before going further. In this lesson, integration means that an AI tool is connected to another app or service, so the AI can see, use, or act on information from that other app without you having to copy and paste it back and forth yourself. Claude's Projects and file uploads, which you learned about earlier in this course, are a related idea on a smaller scale: you upload a file, and Claude can read and refer to it during that conversation. Gemini's integration with Gmail, Google Docs, and Google Drive works on a larger scale, because Google already owns all four of those products, Gemini included, so it can build a much deeper, more direct connection between them than a company connecting to another company's separate product.

Consider Gmail first. A common, practical use of Gemini's Gmail integration is asking Gemini to help you with an email you are looking at or composing, for example asking it to draft a reply to a message in your inbox, to summarize a long email thread so you do not have to read every message in it individually, or to shorten a draft you already wrote before you send it. Because Gemini can be connected directly to your Gmail account, in principle it can work with the content of a specific email without you needing to copy that email's text out of Gmail and paste it into a separate chat window, the way you might have to with an AI tool that has no direct connection to your email at all. For a blind Gmail user managing a large inbox, that kind of direct connection can meaningfully cut down on repetitive copying, pasting, and app-switching.

Google Docs integration works on a similar principle. If you are working in a Google Doc, whether it is one you wrote or one someone shared with you, Gemini's integration is generally intended to let you ask questions about that document's content, request a summary of a long document before reading it in full, or ask for help improving a specific paragraph, without leaving the document or manually copying the whole thing into a separate chat. This can be especially useful for a screen reader user reviewing a very long document, such as a lengthy report or a shared document with many collaborators' comments and edits scattered throughout it, since asking for a summary or a specific answer can be considerably faster than reading the entire document from top to bottom with Say All every time you need one piece of information from it.

Google Drive integration extends this same idea to your files generally. Drive is Google's file storage service, holding documents, spreadsheets, presentations, and other files, often organized into folders. Gemini's Drive integration is generally intended to let you ask Gemini to find a specific file by describing it, rather than remembering its exact name or folder location, or to ask for a summary of a file's contents without opening and reading through it directly yourself first. For a screen reader user, searching a large, unfamiliar Drive folder structure by arrowing through file after file can be slow, so being able to describe what you are looking for in plain language, such as "find the budget spreadsheet I edited last month," and letting Gemini search on your behalf, can save real time, provided you always verify what it finds using good cross-checking habits before relying on it for something important.

Picture a concrete scenario to see why this matters. Imagine you receive a long email thread with a dozen replies about a schedule change, and a linked Google Doc with the full updated schedule attached. Without integration, you would need to open the email, read the whole thread with Say All to find what changed, open the linked Doc separately, read through it as well, and then mentally combine what you learned from both. With genuine integration available and working, you could instead ask Gemini directly to summarize what changed across both the thread and the document in one response. That is the practical time savings integration is meant to offer, and it is the specific advantage this week asks you to notice and evaluate for yourself in the exercises ahead.

It is worth being honest about one thing here, following the same accuracy principle used throughout this course: exactly how Gemini's Gmail, Docs, and Drive integration is presented on screen, which menu, icon, or button surfaces it, and exactly what it is called, can change over time as Google updates these products, and this lesson cannot promise today's exact wording will still match what you hear when you try it. What stays reliable is the underlying idea, integration meaning a direct connection between Gemini and another Google app, and the exploration method for finding it: use your screen reader's list of elements, heading navigation, and button navigation to look for any menu, icon, or option that mentions Gmail, Docs, or Drive by name whenever you are inside Gemini.

A general screen-reader workflow for Gemini

Bringing together everything from this week's navigation and integration sections, here is a general, reliable workflow for using Gemini with a screen reader, built entirely from exploration skills rather than specific memorized steps that could go out of date. First, open Gemini, either the website or the mobile app, and give the page or screen a moment to finish loading before exploring, since AI chat interfaces sometimes load their controls slightly after the page first appears. Second, use your screen reader's list of elements, or your platform's equivalent overview feature, to get a quick sense of what is on the page: how many headings, buttons, and edit fields it contains, before diving into reading or typing anything.

Third, locate the message box using your screen reader's edit-field navigation, or by Tabbing through the page's controls, and type or dictate your question or request. Fourth, submit your message, usually with Enter from inside the message box or by locating a send button with button navigation. Fifth, once Gemini responds, move to the start of the new response using heading navigation, landmark navigation, or by continuing forward through the page from where you submitted your message, and use Say All, or your screen reader's continuous reading command, to hear the full response read straight through without needing to press a key after every line. Sixth, if you are looking specifically for an integration-related control, such as anything connected to Gmail, Docs, or Drive, return to your screen reader's list of elements and scan its buttons and links for anything mentioning those names.

This six-step pattern is intentionally general, because it is built to survive interface changes that are outside this course's control. If a future version of Gemini rearranges its buttons, renames a menu, or changes exactly where the message box sits on the page, every one of these six steps still works, because none of them depend on a specific button being in a specific place. This is the same durable approach used earlier in this course for Claude and ChatGPT, and it is worth explicitly noticing, by this third tool, that you are no longer learning a brand new skill each time; you are applying one core exploration skill, built early in this course, to a third interface.

If something along the way seems to go missing, such as a response that does not seem to load or a control your screen reader's list of elements does not show yet, the same patience-and-retry approach from earlier in this course applies here too: wait a few seconds for the page to finish updating, then repeat the elements-list or heading scan rather than assuming the feature does not exist. Modern chat interfaces frequently update small parts of the page after your action, and your screen reader's browse mode sometimes needs a moment, or a manual refresh of its view of the page, to catch up with what has changed on screen.

Comparing Gemini to Claude and ChatGPT

You have now spent hands-on time with three of the most widely used AI chat tools: Claude, ChatGPT, and Gemini. None of these three tools is simply "the best" one in every situation. Each has genuine strengths and genuine weaknesses, and the right choice depends on the specific task in front of you. This section compares all three honestly across several dimensions that matter in daily use: general conversation quality, file and app integration, mobile experience, and how easily each fits into an existing workflow you might already have. Treat this comparison as a starting point for your own judgment, not a final ranking, since your own hands-on experience in this week's exercises and portfolio project matters more than any general description in this lesson.

On general conversation quality, drafting, explaining, brainstorming, and answering questions, all three tools are strong, capable general-purpose assistants, and for many everyday tasks, such as drafting an email, explaining a confusing term, or summarizing a short piece of text, you may find the differences between them are genuinely small. Claude is often noted for careful, thorough writing and for its Projects and artifacts features, which let you build and revisit larger, organized pieces of work over time. ChatGPT is often noted for its wide range of specialized GPT models and for voice mode as a particularly strong accessibility feature for hands-free, screen-free conversation. Gemini's comparable strength, as this week has emphasized, is not a different flavor of conversation so much as its deep connection to Gmail, Docs, and Drive.

On file and app integration specifically, this is where the three tools diverge the most. Claude's file uploads and Projects let you bring outside documents into a conversation and organize related work together, but that connection generally has to be built manually each time, by uploading a file yourself. ChatGPT's file handling works in a broadly similar, upload-based way. Gemini's integration with Gmail, Docs, and Drive is a different kind of connection: because Google owns all of those products alongside Gemini, the connection is closer to built into your everyday accounts rather than something you set up fresh in each conversation, at least once integration is turned on and available on your account. If your day-to-day work already lives largely inside Gmail, Docs, and Drive, that difference can matter a great deal in practice, saving real time you would otherwise spend uploading files or copying text by hand.

On mobile experience, all three tools offer a dedicated app with screen reader support through VoiceOver on iPhone and TalkBack on Android, and the core reading and typing experience is broadly similar across all three once you are inside the app. Gemini's mobile situation is somewhat different because of its role as a built-in assistant on some Android phones, discussed earlier in this lesson, which means some Android users may reach Gemini through the phone itself rather than only through a separate downloaded app. Claude and ChatGPT remain separate apps you open directly on both iPhone and Android. None of these differences make one tool more or less accessible than another in a general sense; they simply change how you reach the tool in the first place.

It is also worth thinking briefly about privacy and account considerations when choosing between the three, extending the honest, verify-everything habit from earlier in this course. Because Gemini's integration works by connecting to your Google account, using it with Gmail, Docs, or Drive generally means granting Gemini permission to work with data in that account, a genuine convenience with a genuine tradeoff worth understanding rather than accepting automatically. Claude's and ChatGPT's file uploads, by contrast, generally only involve the specific file you choose to share in that moment. Neither approach is simply right or wrong; each is a different balance between convenience and how much account access you are comfortable granting, and that balance is a personal decision worth making deliberately rather than by accident.

Given these honest differences, how might you choose a tool for a specific real task? For a task deeply tied to your Gmail inbox, a Google Doc, or a file already stored in Google Drive, Gemini's integration may save you real time compared to manually copying content into Claude or ChatGPT first. For a task centered on a long, structured writing project you want to build and revisit over multiple sessions, Claude's Projects and artifacts may fit especially well. For a task where you want to speak your request out loud and listen to a spoken reply with minimal screen interaction, ChatGPT's voice mode may be the smoothest experience. None of these are rules; they are starting points for your own judgment, and this week's portfolio project asks you to build your own comparison from direct, hands-on experience with all three tools.

Key terms from this week

Use this list to review the vocabulary introduced in this lesson before you start the exercises and the test.

Gemini

Google's large language model chatbot, connected to Google's other products, including Gmail, Google Docs, and Google Drive.

Integration

A direct connection between an AI tool and another app or service, letting the AI use information from that app without manual copying and pasting.

Built-in assistant

An AI tool made available directly within a device's operating system, such as Gemini on some Android phones, rather than only as a separate downloaded app.

Keyboard-only exercises

These exercises use only your keyboard. General screen reader features are described by what they do; check your own screen reader's shortcut list for the exact keys. Where VoiceOver or TalkBack work differently, that is noted separately.

Exercise 1: Locate Gemini's message box and have a short conversation

1. Open your web browser and go to Gemini's website, or open the Gemini mobile app if you prefer to work on your phone.
2. Use your screen reader's list of links or interactive elements (sometimes called an elements list) to look over the page's headings, buttons, and edit fields before doing anything else. On VoiceOver, use the rotor to review headings and form controls; on TalkBack, explore by swiping through the screen.
3. Use your screen reader's edit-field or form-field navigation to locate the message box.
4. Type a short question, such as asking Gemini to explain a simple topic in two or three sentences, then submit it with Enter or by locating the send control with your screen reader's button navigation.
5. Once a response appears, move to the start of it using heading or landmark navigation, then read the full reply with Say All, or your screen reader's continuous reading command.

Exercise 2: Explore for a Gmail, Docs, or Drive integration control

1. While still inside Gemini, open your screen reader's list of elements again, and look through the list of buttons and links for anything that mentions Gmail, Docs, or Drive by name.
2. If you find such a control, try one simple integrated task, such as asking Gemini to summarize a specific email, a specific Google Doc, or to find a file in Drive by description.
3. If you do not find an integration control, that is a legitimate and useful finding on its own. Note what you searched for and where, and consider whether the feature may require it to be turned on separately for your account, following this week's lesson on how these features can vary.
4. Either way, write two or three sentences describing what you found, using the same honest, exploration-based approach this week's lesson modeled.

Exercise 3: Compare Gemini's answer to an earlier tool's answer

1. Recall a real question you asked Claude or ChatGPT in an earlier exercise in this course.
2. Locate Gemini's message box again using the skills from Exercise 1, and ask Gemini that exact same question.
3. Read Gemini's response in full with Say All, or your screen reader's continuous reading command.
4. Compare Gemini's response to what you remember, or saved, from Claude's or ChatGPT's answer to the same question, noting at least one similarity and one difference.

Portfolio project: My Three-Tool AI Comparison

This week's portfolio piece is a short written comparison of Claude, ChatGPT, and Gemini, based on your own hands-on experience with all three tools.

1. Open Notepad or your word processor of choice.
2. Write a heading at the top: "My Three-Tool AI Comparison."
3. List all three tools: Claude, ChatGPT, and Gemini.
4. For each tool, write two or three sentences describing your own experience navigating it with your screen reader, based on your experience with each tool so far in this course.
5. Pick two or three specific real tasks, such as drafting an email, summarizing a document, or researching a topic, and write which tool you would choose for each one and why, drawing on this week's comparison section.
6. Include a short note on what you found, or did not find, during Exercise 2's search for Gmail, Docs, or Drive integration.
7. Save the file with the name my-three-tool-ai-comparison, in the same portfolio folder you have used throughout this course.

Weekly test

Answer key

Each answer below includes a one-sentence explanation drawn from this week's lesson.