OpenAI、AppleのChatGPT統合の不出来に「裏切られた」と関係者 OpenAI feels “burned” by Apple’s crappy ChatGPT integration, insiders say

OpenAI staff reportedly feel "burned" by how Apple integrated ChatGPT into its devices, according to insiders cited in recent reporting. More than a year after the partnership was unveiled with great fanfare, the real-world experience appears to fall well short of what either side hoped for.

At WWDC 2024, Apple announced that ChatGPT would plug into Siri and system-wide writing tools as part of its broader Apple Intelligence push. Users could escalate tougher questions to OpenAI's model or have it analyze images and documents. In practice, the hook-up is shallow: Siri must explicitly ask permission before each handoff, there is little shared context, and ChatGPT remains an optional sidecar rather than a deeply embedded assistant. For OpenAI, which had hoped to ride Apple's roughly billion-strong active iPhone base into mainstream consumer adoption, the channel has so far underdelivered.

Compounding the frustration is Apple's delay of the long-promised "more personal" LLM-powered Siri, which has slipped into 2026 or later. Apple Intelligence as a whole has drawn mixed reviews, with notification summaries pulled after embarrassing errors and rollout pacing slowed in several markets and languages. Each delay narrows the surface area where ChatGPT might shine on Apple hardware.

The competitive context sharpens the sting. Google has woven Gemini deep into Android and Pixel, struck distribution deals with Samsung, and is steadily replacing Assistant. Microsoft owns the Windows desktop with Copilot and continues to bankroll OpenAI even as the two increasingly compete. Meta is pushing its own assistant across WhatsApp, Instagram, and Ray-Ban smart glasses. Against that backdrop, Apple's cautious, privacy-first integration looks especially constrained, and OpenAI's window of advantage on iOS may be smaller than it expected.

It is not surprising, then, that OpenAI has been investing heavily in its own distribution. The company hired former Apple design chief Jony Ive to develop a dedicated AI hardware device, has rolled out a ChatGPT-branded browser and search experience, and continues to push its mobile apps and desktop clients. If Apple cannot or will not turn ChatGPT into a first-class citizen on the iPhone, OpenAI has clear incentives to build channels it controls outright.

Apple faces its own dilemma. Its brand rests heavily on on-device processing and strict privacy, which makes routing user data to a third-party cloud model philosophically awkward. Yet to keep pace with rivals, it needs capable generative features that its in-house models do not yet match. Reports have repeatedly suggested Apple is also negotiating to add Google's Gemini, and potentially Anthropic's Claude, as alternative back ends in Siri. That would dilute ChatGPT's privileged slot and turn Apple into more of a neutral broker between frontier labs, a position that may serve Apple's interests but further frustrate OpenAI.

None of this means the partnership is in immediate danger. ChatGPT integration is still a marquee feature for Apple Intelligence, and Apple's reach remains unrivaled in premium consumer hardware. But the reported friction is a reminder that in the current AI race, distribution is as contested as model quality. Whoever controls the default assistant on the phone in your pocket holds enormous leverage, and the tension between platform owners and model providers is only likely to grow as both sides try to capture more of the value generated by everyday AI use.