Using AI Mode Within Google Search

In today’s episode, we'll explore the features and functionality of AI Mode in Google search. Visit AVID Open Access to learn more.

Transcript

Paul Beckermann 0:01 Welcome to Tech Talk for teachers. I'm your host, Paul Beckermann.

Transition Music with Rena's Children 0:05 Check it out. Check it out. Check it out. What's in the toolkit? Check it out.

Paul Beckermann 0:16 The topic of today's episode is using AI mode within Google Search.

Paul Beckermann 0:22 By now, most people have used ChatGPT or Google Gemini to generate answers or ideas. The conversational give and take feels intuitive and can be used with very little training or guidance. Now, gradually, much of that functionality is finding its way into the world's most popular search engine, Google Search, and that functionality is continually evolving to bring in more AI features.

Most people have probably noticed the AI overviews that appear at the top of a traditional Google search. You could type in a prompt, such as "examples of Robert Frost's use of imagery in his poetry," click enter, and Google will give you an accessible summary at the top of the page. This gives you a quick, scannable response to your prompt. To the right of that overview summary is a list of links that were used to generate the displayed summary. This is really important and helpful, as it allows you to check on Google's work.

While the overviews are generally accurate, they can be wrong. Clicking into the source links lets you not only check for accuracy, but also dig deeper into the content Google is referencing in its summary. If you want even more, you can scroll down the page for a more traditional list of website links produced by your prompt.

If you're looking for quick answers or questions that are not critically important, this process can save lots of time. Google essentially acts as your research assistant by scanning relevant web pages and then using AI to generate the summary. If you're looking for answers of high importance that must be correct, you'll need to proceed with more caution. In this case, you can still read through the overview summary to acquaint yourself with a more in-depth understanding of the topic before proceeding to a deeper dive into that list of links that has been generated. Even while digging into those links, you still need to vet out those sources for credibility and accuracy. Both pathways require your human skills of content verification and credibility detection. With that context in mind, in today's episode, I'm going to explore the newest AI interoperability built into Google searches. It's called AI mode. It was introduced in the spring of 2025 and has been gradually expanding globally since.

Transition Music with Rena's Children 2:35 Here's the tool for today.

Paul Beckermann 2:39 To access AI mode, log into your Google account and then open up Google Search and click the "AI mode" button at the right of Google's search field, where you type in your query. Upon clicking that, you'll be greeted by a page that welcomes you with the heading, "Meet AI mode," and then the sub caption, "Ask detailed questions for better responses." The search box here cues you to ask anything, and there are some sample searches listed below based on your past browsing history to get you started. There are also options to click a microphone and speak your prompt or click the plus icon to upload an image or file, and even an option to create an image using Nano Banana.

In a sense, you're getting Google Gemini generative AI embedded into Google Search. Input a prompt, and then click enter. You'll get an AI-generated summary similar to what you would get with the regular Google AI overview, and you'll also get a list of sites used to generate that summary.

So you may be asking, "How is this different from the AI summaries you get in a regular Google Search mode? Why shouldn't I just use that?" Well, Google describes the difference this way: AI overviews and AI mode are two distinct generative AI features in Google Search that differ in their integration, interaction style, and the complexity of queries they handle. AI overviews are automatic summaries within standard search results, while AI mode is a separate interactive and conversational experience for deeper research.

Let me break that down further into six specific differences.

The first difference is the interface itself. AI overviews appear as a summary at the top of a traditional search results page, while AI mode replaces the entire results page with a chat-like conversational interface.

The second difference includes how a user triggers the feature. With AI overviews, this is automatic for queries where Google's system determines a summary would be helpful. AI mode, on the other hand, is user-initiated and accessed by clicking the "AI mode" button or by navigating to google.com/aimode. This is considered better for complex, multi-layered queries.

Number three, this difference involves the user interaction. AI overviews offer a single, static snapshot of information with source links. Additional interaction questions require clicking on individual links. In AI mode, the interaction is more dynamic and conversational, allowing users to ask follow-up questions in a back-and-forth dialogue, carrying context from previous questions to subsequent inquiries.

Number four addresses query type. Traditional searches are best for straightforward questions, definitions, and quick fact-finding, where a concise answer is what is desired. AI mode has been designed for more complex tasks, research, planning, and comparisons that require deeper explorations, reasoning, and follow-ups.

Number five concerns technical capacity. AI overviews use a custom Gemini model to provide a summary of one round of AI analysis. AI mode, in contrast, uses a more powerful custom system of Gemini and the query fan-out technique to perform multiple parallel searches and provide a more comprehensive answer. Essentially, AI mode divides a complex query into subtopics and then searches them simultaneously, synthesizing a response from those results.

And number six is availability. AI overviews are widely available to users all across the globe, while AI mode is being rolled out, starting with the U.S., and requires users to be signed into their Google account.

In many ways, AI mode is like using the Gemini generative AI chat bot with traceable web citations included in the interactive experience. I should note that sources can generally be listed in the chat bot as well, but with AI mode, you get those by default, and it's initially accessible from a Google Search screen, so it may save you some time.

Here are two bonus and probably more advanced ways that you can use AI mode.

Transition Music with Rena's Children 6:47 Here's a bonus tip.

Paul Beckermann 6:50 Number one is multi-modal input. Chrome's AI Mode supports using text, voice, or image inputs. You can even use the integrated Google Lens inside of Chrome to ask the AI about what you're seeing on a page. For instance, you might have a spreadsheet open, and you can ask the AI where to click to add a new row. Similarly, you could upload images and ask questions about them.

And the second bonus tip is cross-tab analysis. If you have multiple tabs open in your browser about a topic you're researching, you can ask the AI to compare tabs or summarize findings across those tabs. This can save you lots of time bouncing back and forth from tab to tab.

Paul Beckermann 7:29 All right, I want to wrap up this episode by a little discussion about using AI mode to teach, rather than just getting answers. One concern I often hear about generative AI is that it can shortcut student thinking, simply giving students an answer rather than making them think. This is a valid concern. In response to this, I tried out a strategy to turn AI mode into a bit of a tutor. Here's what I did.

I began with some AI questions about Robert Frost's use of metaphors in his poetry. The AI gave me a great summary, as well as specific examples from his work. It also gave me a list of links where I could dig deeper. Then I decided to take it a step further into application. I prompted the AI to teach me how to write a metaphor in the style of Robert Frost. It gave me a nice overview, including key characteristics of Frost's metaphors. It followed that up with a step-by-step guide to writing a Frost-style metaphor, and the five actionable steps included examples for each step. It also gave me a more complete final example with analysis.

In response to that, I followed the steps up by providing my own metaphor in the style of Frost—one that I wrote. I pasted it into the chat box and asked the AI to give me feedback on my original example. This is where I felt like the real learning happened. The AI responded as a kind teacher, giving me five areas of strength in my example. It then offered what it called an "optional minor area of refinement," a nice way of saying, "I could improve it by doing this." It gave me a nice suggestion for rewriting, sharing that while I used a classic metaphor structure that explicitly compared a barn door to a mood, I could have mimicked Frost's style more directly by implying that comparison, not stating it explicitly.

In a scenario like this, a student could rewrite the example and re-upload it into the prompt, asking for more feedback on that revision. In this type of interaction, I feel like the AI is acting more like a tutor or coach than an answer engine.

I then put on my teacher hat, and I asked the AI to turn this thread into a lesson plan to help ninth-grade students understand and practice metaphor writing in that distinctive style of American poet Robert Frost. In response, the AI gave me a really nice, quick lesson plan that I could integrate almost immediately with my students. If there were any portions that I didn't really care for, I could ask follow-up questions or suggestions and have the AI revise the lesson, perhaps with more group work, for example.

Paul Beckermann 10:04 In short, the auto-generated AI overviews in a normal Google Search can be very helpful for getting quick answers to simple questions. If you're looking for more in-depth conversation with a thought partner, AI mode might be the better choice. It gives you much of the same experience you'd get from a full Gemini chat bot with the convenience of accessing it through Google Search and with auto-generated source links as well. I'd suggest giving it a try and seeing how it works for you.

Paul Beckermann 10:31 To learn more about today's topic and explore other free resources, visit Avid Open Access.org. Specifically, I encourage you to check out the article collection, "AI in the K-12 classroom," and of course, be sure to join Rena Winston and me every Wednesday for our full-length podcast Unpacking Education, where we're joined by exceptional guests and explore education topics that are important to you. Thanks for listening. Take care and thanks for all you do. You make a difference.