By Philip Michaels
May 21, 2026 9:20 AM PT
Google I/O: What you need to care about when you don’t care about Google I/O
Instead of attending this week’s Google I/O keynote — an event I’ve been covering for the past decade — I found myself in a physical therapist’s office getting work done on an arm I broke last year. While Sundar Pichai was outlining the latest developments in AI and how they’ll affect each and every one of Google’s products, a polite-yet-determined physical therapist was busy yanking my arm into all sorts of positions, with the hope of stretching it back into its regular shape.
I think I got the better end of the deal.
OK, that’s somewhat unfair to Google, which had plenty to announce at its annual developer conference, with a lot of it impacting software and services you likely use. To be sure, there was plenty for developers and coders to sink their teeth into — only at Google I/O are you going to get applause breaks when a speaker mentions the number of tokens Google is processing or the speed of tensor processing units. But Google made announcements of interest to civilians like you and me as well.
Don’t believe me? Well, if you’ve ever used Google’s search tool, the company has some big changes in store. And with 13 Google products having at least a billion users — and some like Gmail and YouTube pass the 3 billion user mark — there’s a good chance you’re a member of Google’s customer base, whether you want to admit it or not. Changes are coming there, too.
But will they be positive changes? That’s harder to discern at this point, though Google is certainly enthusiastic. “Let’s make something that matters,” a hype video kicking off the nearly 2-hour keynote intoned. Given the ambivalence that some of us feel toward the inclusion of AI in just about everything and the current direction of travel for the tech industry, the jury is still deliberating over whether Google succeeded in that goal.
For anyone who doesn’t live and breathe Google pronouncements, here’s a quick rundown of the Google I/O developments you should file away for later reference.
A new approach to search
If you’ve been on the web long enough, you likely think of Google Search results as a ranked list of clickable links that ideally point you toward whatever it is that you’re looking for. Don’t get too comfortable with that idea, though, as Google is changing the look of its search tool, and I’m not just talking about the updated search box which dynamically expands to fit larger queries and taps directly into Google’s latest AI tools.
In other words, the AI Mode currently available in Google Search sounds like it’s going to be more of the default view, with queries producing a detailed summary of what the search engine found, along with prompts to pose follow-up questions that drill down into the results. In theory, that makes search a more conversational process that better delves into the nuance of what you’re specifically looking for. In the coming months, agentic coding is going to adapt the look of search results to reflect their content, and Pro and Ultra subscribers will be able to put agents to work searching for things like vacation rentals and newly released sneakers.
But before all that, the problem remains that AI Mode searches can produce some laughable results. I ran an AI Mode search for myself just now and learned that I’m the managing editor at Tom’s Guide — surely, that’s going to surprise the people there who laid me off back in January. (Sorry, boys — Google says I’m back on the payroll!) On the bright side, Google, I do appreciate being described as a “prominent personal technology journalist” — now those are the kind of hallucinations I can get behind. A follow-up query for my social media profiles included my Twitter handle, which I deactivated and deleted more than a year ago; not to worry, though, as the link takes you to an entirely different dude named Philip Michael.
You can understand, then, why the news that Google is going to offer more of this makes me a little nervous. Summaries are nice, but they’re not very helpful time savers if I have to fact-check each piece of information. And with Google searches driving even less traffic to some sites, how long before publishers start moving more things behind a paywall, resulting in even more incomplete results?
We’re not going to find out the answer to that question until we have a chance to use Google’s revamped search tool, which is in the process of rolling out. Get ready for a summer of learning how to search all over again.
More conversations in more products
Gemini found its way into Google Maps back in March, when Google added an Ask Maps feature that let you perform more complex searches. (Think “I’ve got a meeting in half-an-hour and I need to eat; where’s a quick place to get a bite to eat that’s a short walk from here?“) A whole host of other Google Apps — we’re talking YouTube, Gmail and Google Docs — are now following suit, with AI-powered features on their own.
The new conversational search capability makes sense for YouTube, where you’ll be able to make more specific requests for the types of videos you’re looking for. Presumably, this feature will be especially handy if you’re looking for how-to videos for handling a very specific task.
I’m less convinced about Gmail Live, a voice-powered feature where you’ll be able to ask Gemini for specific information — drop-off details for school field trips, flight departure info, the details about a party you’ve been invited, too — and your assistant will find what you’re looking for and read it back to you. We’ll see how that works in practice, but anything that improves on Gmail’s frustrating search capabilities will be a welcome addition. (AI Inbox in Gmail is getting new capabilities, too, at least if you subscribe to Google AI Plus or Pro. Here, the updates promise to generate contextual drafts for emails that require a prompt reply as well as surface Docs, Sheets and Slides links that require your review.)
A new Docs Live feature, though, represents everything I hate about AI. Google bills this as a way to help you get started with writing projects — you just brainstorm ideas and Docs Live turns them into a draft for you to review and refine. A Docs Live demo video posted by Google has a software engineer putting the AI agent to work formulating talking points for a Career Day presentation, going so far as to task Docs Live with producing some funny analogies on his behalf.
Google is taking pains to frame this tool as a way for getting started on writing projects, but you and both know that less scrupulous users are going to offload writing entirely on to Docs Live. You could argue that none of this is Google’s problem, as it’s just handing people a tool, and it’s up to them to use the tool responsibly. But that’s sort of like handing out booze and heavy machinery and expecting everyone to mind their Ps and Qs. Sometimes, you have to be more careful with the kind of work you’re enabling.
New creative tools
Google touted a few generative media tools — some new, some updates to existing features — and the most relevant one to our purposes is Google Pics, an image creation tool coming to the Google Workspace this summer. With Pics, you be able to edit photos, create posters and flyers and whip up illustrations to share on social media, using text commands.
Pics tools that Google showed off during its I/O keynote include moving selected images around, resizing them, and getting unwanted distractions out of photos. You’re also able to integrate text into mages, even changing around the text that’s already in a photo.
A lot of those image-editing tools will sound familiar to anyone who’s used similar features to play around with photos on a Pixel phone, and those have been pretty polished in my experience. A watermarking tool will let people know that the image has been manipulated to head off any funny stuff like deep fakes.
Gemini Spark and Gemini Omni
Two other Google I/O AI announcements deserve some attention. Gemini Spark is an AI agent coming to the Gemini app for business customers, though you imagine it will eventually roll out elsewhere. Spark integrates with Google’s products as well as some other third-party tools to handle tasks on your behalf, such as generate emails, pulling details from reports and monitoring changing information online.
Google built Spark so that it’s working 24/7 — it’s a cloud-based assistant — which means that it’s still monitoring things even when you’ve shut down your laptop or put your phone to sleep. I’d have to see it in action beyond an on-stage demo to get a sense for how it performs though Google stresses that Spark will seek your permission before it acts on anything.
Gemini Omni is the last bit of big software news to come out of Google I/O. According to Google, the ultimate goal with Omni is to be able to create anything from any input, which certainly sounds ambitious. At the moment, Omni’s focus appears to be on video, with Google promising that you’ll be able to generate a video from any text, image or audio. The first iteration of Gemini Omni — Gemini Omni Flash — is part of the Gemini app, Google Flow and YouTube shorts.
The Omni demos shown off by Google took videos shot by people, using voice inputs to change things like the video’s style, environment and even viewing angles. In the hands of a creative pro, it could be quite a powerful part of an already advanced skill set; used by less skilled practitioners, it could increase the amount of AI slop we see out in the wild. But then, that warning could apply to a lot of the AI capabilities Google is adding over the next few months.
Google’s glasses (but not Google Glass)
Finally, we saw some hints of hardware at this year’s Google I/O. Specifically, Google showed off some new smart glasses built on its Android XR platform, with the likes of Gentle Monster and Warby Parker supply the designs. The glasses will come in audio-only versions, where a Gemini assistant talks in your ear to you (and only you), while a display version of the glasses will info right in front of you.
During its keynote, Google touted the hands-free nature of the experience, with the on-board assistant able to give you directions, help you send texts and emails and identify whatever it is you’re looking at — all without having to take your phone out of your pocket. But you will need to be carrying a phone to pair with your glasses — Google says they’ll work with iPhones as well as Android devices — which makes me wonder what the point is.
Sure, we could all spend more time not staring at phone screens. But is that worth the extra money you’ll have to spend on glasses for an accessory to a phone that already costs hundreds of dollars on its own? (For what it’s worth, there’s no pricing information on these Google glasses yet, as they’re not arriving until later this fall.)
If you remember, the Apple Watch debuted with the same premise — it’s really just an extension of your phone. And while that selling point may have brought in a few early adopters, the Apple Watch itself didn’t take off as a product in its own right until the health and fitness tracking features really came to the fore. I think mixed reality glasses — whether it’s these Google-backed models or the pair that Apple is allegedly developing — are going to have to carve out their own space beyond just “smartphone stuff but the smartphone stays in your pocket.”
Maybe something to focus on for I/O next year.
[Philip Michaels has been writing about technology since 1999, most notably for Macworld and Tom’s Guide. He currently finds himself between jobs, so if you need someone who can string a few sentences together (or make your sentences read a lot better), drop him a line.]
If you appreciate articles like this one, support us by becoming a Six Colors subscriber. Subscribers get access to an exclusive podcast, members-only stories, and a special community.