Your iPhone Just Became a Chatbot Ballot. Pick Carefully.

There is a settings menu coming to your iPhone this fall. You will find it buried a few taps deep inside Apple Intelligence and Siri. The question it asks is deceptively simple: which AI do you want?

Claude. ChatGPT. Gemini. Grok. Take your pick.

At WWDC on June 8, Apple announced that iOS 27 will include what it calls "Intelligence Extensions" — a framework that lets third-party AI providers plug into Siri, Writing Tools, and Image Playground. The first wave includes Claude, Gemini, and Grok alongside ChatGPT. Not defaults. Options. You choose.

The tech press framed this as Apple "opening up its platform." That is too generous to Apple and too dismissive of what it actually signals.

The Browser Wars Rhyme

If you were online in 2009, you remember the browser ballot screen. Microsoft, under pressure from European antitrust regulators, had to add a window to Windows that asked users to choose between Internet Explorer, Firefox, Opera, and Chrome. The whole thing felt slightly humiliating for Microsoft — an admission that their browser monopoly was a commercial decision, not a user preference or a technical truth.

Apple did not have to do this. No regulator forced iOS 27's Intelligence Extensions framework into existence. Which means something more important happened: Apple read the market and decided the era of picking your users' AI for them was over.

They were not wrong.

The Numbers That Explain Everything

On June 16, Sensor Tower released its State of AI 2026 report. The headline: for the first time since ChatGPT launched in November 2022, no single AI chatbot commands more than half the market. ChatGPT's share fell to 46.4% by May 2026. Gemini stands at 27.7%. Claude accounts for 10.3%. The rest is distributed across a growing field.

Three and a half years of dominance, and now: a market.

This is not a story about ChatGPT declining. It crossed one billion monthly active users. That is not a product in trouble. That is a product at scale in a market that grew faster than any single company could contain. The number that matters is not that ChatGPT slipped. It is that the combined gains of Gemini, Claude, and other alternatives pulled enough market share to break the majority. Users diversified before the platforms even asked them to.

Apple noticed. A settings menu for AI is only worth building if users care which AI they are using. It turns out they do — and they are already voting with their usage patterns.

What You Actually Choose When You Choose

Here is the part that is less comfortable to say: the AI you reach for first is not a neutral utility decision anymore.

ChatGPT is OpenAI, which is now entangled with the U.S. Department of Defense, SoftBank, and a proposed $500 billion national infrastructure program. Gemini is Google, which rebuilt its search results using AI and triggered a spike in DuckDuckGo installs from users who did not like what they found. Claude is Anthropic, which navigated federal pressure to lock down its most capable models while publishing responsible scaling commitments that the Vatican, of all co-authors, endorsed. Grok is Elon Musk's operation, which should be self-explanatory to anyone who has been online this decade.

Choosing your default AI is choosing whose infrastructure processes your ideas. Whose training data shapes what gets surfaced. Whose safety filters, or deliberate absence of them, determines what you are allowed to ask.

Most people have not thought about it this way yet. The iOS 27 ballot screen will make some of them start.

What the Chatbot Makers Are Actually Competing For

Being the default AI on two billion iPhones is the highest-value real estate in consumer technology right now. Google reportedly paid Apple somewhere around $20 billion a year for the privilege of being the default search engine on iOS. The AI equivalent of that deal has not been formalized — and the fact that Apple is building an open competitive framework rather than accepting a single exclusive payment suggests the market is too uncertain, or too contested, to pick a winner and bank on it.

That is an unusual posture for Apple. They like picking winners. They picked the music player format, the smartphone format, the tablet format, the wearables format. The fact that they are building a marketplace instead of choosing a champion tells you something: they do not know which AI will be leading in three years. Neither does anyone else.

The companies making bold bets on AI infrastructure are staking enormous capital on particular futures. Apple is hedging. In 2026, that might be the smarter play.

The Choice That Matters

The browser ballot did not change market share overnight. Internet Explorer dominated for years after Europe forced the screen on Microsoft. What changed was the cultural understanding that browser choice was a legitimate preference — that IE's default status was a business decision, not a feature.

The AI ballot screen coming to your iPhone will work the same way. It will make explicit what has been implicit: there is no natural default. Every AI currently in your phone got there because someone chose to put it there. Some of those choices were commercial. Some were technical. None of them were inevitable.

The market share data tells you that people are already making the choice, even without being asked. They are downloading Claude because it is better for writing. They are keeping Gemini because it lives inside Google Docs. They are staying on ChatGPT because it got there first and the switching cost felt high.

This fall, your phone is going to ask you directly. The answer you give will be more meaningful than a settings selection. It will be the first time most people treat AI as a preference rather than an inevitability.

Use that carefully. The AI you choose is not just a tool. It is an infrastructure decision. And infrastructure decisions, once made, tend to compound.