Google just spent two hours at I/O 2026 describing a future where you never have to do anything yourself again. The search box will expand as you type. Gemini Spark will organize your events. Daily Brief will tell you what’s coming up. An AI will read your Gmail, write your replies, and build you to-do lists. All of it seamless. All of it helpful. All of it happening inside the friendly blue Google logo we’ve trusted for 28 years.

And Demis Hassabis, the CEO of Google DeepMind, closed the keynote by saying we’re standing in “the foothills of the singularity.” A profound moment for humanity, he called it.

Here’s what I think: Google is trying to become the operating system for your entire life, and we should be way more skeptical about that than the I/O audience seemed to be.

The everything box

Let’s start with what Google actually announced. The search bar, that simple rectangle that’s barely changed since 1998, is becoming something fundamentally different. It’s not just for finding information anymore. It’s becoming an interface for AI agents that monitor things in the background, proactively alert you to updates, and take actions on your behalf.

This isn’t search. This is delegation.

Google wants you to ask it to handle your travel plans, manage your schedule, track your packages, and apparently just… live your life for you. The pitch is frictionless convenience. Type what you want, get it done. No need to open multiple apps, compare options, or think too hard about anything.

The problem is that “doing everything” requires knowing everything. And that’s where this gets uncomfortable.

The data you’re not thinking about

Google was refreshingly honest about one thing at I/O: these features require massive amounts of personal data. Your emails. Your calendar. Your location history. Your browsing patterns. The content of your conversations. All of it fed into models that will make decisions about what you see, what you do, and how you spend your time.

The Verge’s coverage pointed out that many of these features depend entirely on your trust. That’s true, but it understates the problem. It’s not just about trusting Google to keep your data secure. It’s about trusting them to use it well, to not manipulate you, to not optimize for their revenue over your actual interests.

And here’s the thing: Google’s entire business model is still advertising. When an AI that reads all your email and knows your entire schedule starts “helping” you make decisions, whose interests is it really serving?

Agents everywhere, judgment nowhere

Google kept using the word “agents” at I/O. Agents that monitor topics for you. Agents that summarize information. Agents that draft responses. The new Gemini 3.5 Flash model is apparently being used for “everything,” according to their own announcement.

But here’s what nobody asked during that two-hour keynote: what happens when you outsource judgment itself?

If an AI is reading your email and deciding what’s important, you’re not learning to prioritize. If it’s drafting your responses, you’re not thinking through what you actually want to say. If it’s organizing your day, you’re not making intentional choices about your time. You’re just… accepting whatever the algorithm surfaces.

This isn’t a theoretical concern. We already know what happens when we let algorithms curate our information diet. We got filter bubbles, radicalization pipelines, and a mental health crisis among teenagers. Now Google wants to give those algorithms even more control, and we’re supposed to be excited about it?

The singularity talk is not helping

When Demis Hassabis started talking about the singularity, I genuinely don’t know if he was trying to inspire people or just got carried away. Either way, it was the wrong note to end on.

The singularity is a concept about artificial superintelligence fundamentally transforming human civilization. It’s speculative, contested, and definitely not something you should be casually name-dropping in a product launch. Especially when the products you just announced are things like “an AI that reads your Gmail.”

It makes the whole presentation feel less like a tech demo and more like a religion. And when a company that already has unprecedented access to human behavior starts talking about profound transformations of humanity, that should make you nervous, not excited.

What we should actually want

Look, I’m not anti-AI. There are legitimate uses for these tools. Summarization can be useful. Translation breaks down barriers. Search genuinely got better with AI-powered understanding of context.

But there’s a difference between tools that augment your capabilities and tools that replace your judgment. There’s a difference between AI that helps you find information and AI that decides what information you should see. There’s a difference between assistance and abdication.

What Google showed at I/O 2026 is mostly the latter category. They want to be the interface between you and everything. They want to handle the boring stuff, sure, but also the stuff that actually matters: what you pay attention to, how you communicate, how you make decisions.

And the cost of that convenience is giving Google even more power over your daily life than they already have. Which is already quite a lot.

The question nobody’s asking

Here’s what I want to know: at what point does “helpful” become “controlling”? At what point does “convenient” become “infantilizing”? At what point do we look up and realize we’ve handed over so much of our decision-making to Google that we’ve forgotten how to do it ourselves?

The I/O keynote didn’t ask those questions. The audience didn’t seem to be asking them either. Everyone was too busy clapping for features that will “help unlock AGI’s incredible potential for the benefit of the entire world,” as Hassabis put it.

But I think those questions matter more than whether the search box dynamically expands or not.

Google is betting that you want them to do everything for you. That you’ll trade judgment for convenience, autonomy for ease, privacy for features. Maybe they’re right. Maybe that’s exactly what people want.

But we should at least be honest about what we’re trading away. Because once you let an AI agent organize your entire life, getting that agency back is going to be a lot harder than typing a query into a search box.

Even one that dynamically expands.