Apple might be rethinking its approach to Siri as the race for the best AI assistant heats up. With Gemini already winning the race, it looks like Apple might take help from OpenAI and Anthropic to power Siri. A new report from Bloomberg claims Apple is exploring deeper integration with AI models developed by OpenAI and Anthropic, marking a significant shift from its usual in-house-first strategy.

Why this Change? 

Apple has been quietly working on its own AI-enhanced assistant, reportedly called “LLM Siri.” But according to the report, the company is now testing OpenAI and Anthropic models on its own cloud infrastructure. This clearly shows that Apple might be struggling to properly train its own AI. Even recently during WWDC, most Apple fans weren’t too impressed with the Apple Intelligence features that were announced. The planned Siri upgrade was originally expected in 2025, but technical delays have now reportedly pushed the timeline back to 2026 or later.

On the other hand, rivals like Gemini, ChatGPT, and Anthropic (Claude) are pushing the limits of their AI systems. This might have forced Apple to join hands with proven third-party AI providers.

Currently, Apple has already integrated OpenAI’s ChatGPT into Siri for select queries like image generation and web searches, but Siri’s core tech still runs on Apple’s own Foundation Models. This could change soon, as Apple has reportedly asked both OpenAI and Anthropic to train custom models to run on Apple’s private cloud infrastructure. These trials are ongoing and are meant to boost Siri’s intelligence without sacrificing user privacy or Apple’s control over the ecosystem. According to Mark Gurman, Claude has been found to be the best in terms of safety and overall performance.

It’s rare for Apple to hand over core functionality to an outside company, but with Siri still lagging behind, this might be a necessary step. Whether it’s OpenAI’s ChatGPT or Anthropic’s Claude under the hood, Siri’s evolution is shaping up to be Apple’s biggest AI bet yet.