Apple is reportedly rebuilding Siri into a fully conversational AI assistant that behaves more like ChatGPT and Gemini, a sweeping redesign that could put a chat interface at the center of the iPhone experience. The effort, codenamed Campos and detailed by Bloomberg’s Mark Gurman, points to a back-and-forth, memory-aware Siri embedded across iOS, iPadOS, and macOS — not a bolt-on chatbot, but a system-level upgrade.
The report suggests Apple will fine-tune Google’s Gemini for on-device intelligence and may lean on Google’s TPU infrastructure for certain cloud tasks, mirroring the hybrid approach Apple already champions for sensitive data. If realized, this would give hundreds of millions of Apple users ChatGPT-class capabilities without leaving the default assistant.
What the New Siri Could Do With Chat-Style Intelligence
The marquee change is a chat-first Siri with persistent context, natural follow-ups, and a clearer understanding of what’s on screen. Think requests like “Summarize this PDF and draft a reply that sounds friendly but direct” or “Explain this chart and add a slide for the key trend,” all handled in one fluid thread.
Apple is expected to fuse this with deeper app control via App Intents and Shortcuts, turning plain-English commands into actions across third-party software. Practical examples: “Send the latest contract to Priya as a PDF, then schedule a recap call for tomorrow” or “Clean up this photo, brighten the subject, and share it in our group,” with Siri orchestrating Files, Mail, Calendar, Photos, and messaging without manual taps.
A chat UI accessible by button or voice would likely live systemwide, capable of rewriting text, generating images or emojis, summarizing notifications, and answering questions tied to personal context — all areas where current assistants fall short compared with modern LLMs.
Why Distribution Changes the Game for Apple’s Siri
Apple’s scale is the x-factor. The company has disclosed more than 2.2 billion active devices globally, and research firms like IDC have placed iPhone at or near the top of worldwide smartphone shipments. In the U.S., Counterpoint Research estimates iPhone holds a dominant share of the installed base.
Contrast that with standalone chatbots. OpenAI has cited more than 100 million weekly active users for ChatGPT, and the mobile app has surpassed the nine-figure install mark, according to data.ai. But default placement beats downloads: if Siri matches the utility of third-party bots and is available from any screen, the extra friction of opening a separate app could become a deal-breaker for many users.
Apple also has little incentive to inject ads into its core assistant experience. While the company sells advertising in select services, its historical playbook favors hardware-led profits and low-friction software. A polished, ad-free assistant would sharpen Apple’s differentiation as rivals explore ads and tiered access.
Under the Hood and the Privacy Question Explained
Technically, a hybrid strategy makes sense. On-device models cut latency, work offline, and keep more data local; cloud models handle heavier lifts. Apple’s recent research — from methods to run larger models off flash storage to multimodal systems that reason over text and images — shows a long-term plan to push as much intelligence to the edge as possible.
The report that Apple could tap Google’s TPUs for cloud inference raises an obvious question: privacy. Apple has emphasized verifiable protections for server-side processing, including designs where user data is processed without long-term retention. If Google’s infrastructure is involved, expect Apple to detail strict technical and contractual controls to preserve those guarantees — anything less would undercut a decade of privacy messaging.
The Competitive Stakes for Google, OpenAI, and Apple
For Google and OpenAI, a deeply upgraded Siri is both threat and opportunity. It could cannibalize standalone chatbot usage on iOS while simultaneously driving massive inference volume to whichever model powers Apple’s stack. If Gemini becomes the default behind Siri’s curtain, Google gains scale that few distribution deals can match.
There’s also a chip subplot. Hosting large-scale inference on Google’s TPUs would nudge demand toward custom accelerators and away from general-purpose GPUs, where Nvidia has enjoyed a surge tied to generative AI. Even a partial shift of consumer assistant workloads could reshape cloud procurement math.
Developers may be the immediate winners. A smarter Siri with richer APIs could finally deliver on the promise of voice-driven app actions, with LLMs handling ambiguity and multi-step tasks. If Apple couples that with clear policies and robust tooling, we may see a wave of “assistant-native” app experiences.
Bottom line: Apple appears ready to turn Siri from a voice command layer into a full conversational agent. If it can pair ChatGPT-grade capability with Apple-grade privacy and reliability, the default assistant on the world’s most influential devices could become the most important AI product of all.
