Apple’s Surprising AI Strategy for Siri Reportedly Includes OpenAI or Anthropic

Apple is reportedly considering partnering with external companies including Anthropic or OpenAI to overhaul Siri’s artificial intelligence, marking a potential departure from its long-standing strategy of building AI technology internally. 

According to Bloomberg’s Mark Gurman, Apple has explored licensing large language models (LLMs) from Anthropic and OpenAI to power a smarter version of Siri. The company has also requested custom configurations that could operate on Apple’s cloud infrastructure to maintain tighter privacy controls.   

If Apple proceeds, this would signal a notable acknowledgement that its internal AI development has struggled to keep pace with competitors including Google and OpenAI. Siri has faced criticism for lagging behind other AI assistants, and Apple’s attempts to upgrade it have repeatedly faced delays.

Why Apple might make the switch

Apple has been developing its own AI systems, known as Apple Foundation Models, with plans to integrate them into Siri by 2026. However, internal evaluations indicated that third-party models — particularly Anthropic’s Claude — outperformed Apple’s tools in handling complex user queries, as reported by Bloomberg.   

Executives including software chief Craig Federighi and Siri head Mike Rockwell have become increasingly receptive to adopting external AI to accelerate improvements. This approach mirrors Samsung’s strategy, which brands its features under the Galaxy AI label while relying on Google’s Gemini models for functionality.

Privacy concerns and internal struggles

Apple is widely known for its focus on user privacy, so any deal with an external AI provider would likely involve deploying those AI models on Apple-controlled servers rather than external cloud platforms. The company has already tested this arrangement in preparation for a possible rollout, noted Gurman.

He also reported that the prospect of outsourcing key parts of Siri’s intelligence has created unease within Apple’s organization; some engineers reportedly fear that bringing in outside models could signal a lack of confidence in their work. Morale has reportedly suffered, with at least one senior researcher, Tom Gunter, recently leaving the company.

Bloomberg reports that Apple has not made a final decision yet and its in-house AI development continues. But if Apple chooses to work with Anthropic or OpenAI, users might see significantly upgraded features as early as next year. 

For now, Apple is moving ahead with some AI upgrades, including allowing third-party developers to use its on-device AI models later this year, according to Gurman.

Editor’s note: This content was originally published on our sister site TechRepublic on July 1, 2025.