OpenAI says ChatGPT features on iPhones saw low user uptake

In short: Court documents say OpenAI was disappointed with how few iPhone users turned on Apple’s ChatGPT features, and Apple has since shifted key Siri work to Google’s Gemini.

What happened

Court documents cited by the Financial Times say OpenAI believed its 2024 deal with Apple did not bring the boost it expected. The deal put ChatGPT into Apple Intelligence, a set of AI features on iPhones.

OpenAI expected a “halo effect,” meaning Apple’s promotion would help more people try ChatGPT and pay for subscriptions. But a month after launch, OpenAI said the rollout started slowly and it lowered its forecasts for weekly users. By summer 2025, OpenAI said the integration was “dramatically underperforming.”

One reason mentioned in the filing was that the ChatGPT option was turned off by default. Users had to switch it on through several steps, like a new feature hidden behind a few extra menu taps.

Since then, the relationship has deteriorated. Apple has partnered with Google to use Gemini models for a rebuilt “Siri AI,” which launched last week in English and is labeled “beta” (a public test version). The companies are also tied up in multiple legal fights, including Apple’s trade secrets lawsuit against OpenAI. OpenAI has also argued that the Apple deal did not unfairly boost its position, and that competition from Google, Anthropic, Meta, and Grok continued to grow.

Why it matters

Many people meet AI tools through the apps already on their phone. This dispute highlights a simple but important point, design choices like whether a feature is on by default can decide whether millions of people ever use it.

Source: Financial Times