Apple Considers Anthropic and OpenAI to Reinvent Siri with Advanced AI Integration

Apple is in advanced talks with leading artificial intelligence firms Anthropic and OpenAI to potentially integrate their large language models (LLMs) into an upcoming version of its voice assistant, Siri. This move marks a dramatic shift in Apple’s AI strategy, traditionally focused on in-house development. By leveraging external AI technology, Apple aims to close the gap between Siri and more capable AI chatbots like ChatGPT and Claude, bringing Siri up to speed in the competitive AI landscape.

The proposed integration would rely on Apple’s Private Cloud Compute servers, running on proprietary Mac chips to ensure high performance and data privacy. Although Apple has previously used its own “Apple Foundation Models” for AI functions, the shift to third-party models would significantly deviate from its original 2026 plan. Internally, testing of these external models has been promising, especially Anthropic’s Claude, prompting further evaluation by Apple’s leadership team.

Siri, introduced in 2011, has struggled to keep pace with rapid AI advancements. Recent internal restructuring has seen Mike Rockwell take charge of Siri engineering, alongside Craig Federighi’s software oversight. Their joint review concluded that third-party AI might better support Siri’s future capabilities, following repeated delays in Apple’s in-house AI projects. This realization signals a more pragmatic approach to meet rising user expectations in the age of generative AI.

However, the proposed shift has sparked concern among Apple’s internal AI teams, who worry it may devalue their work and contribute to a brain drain. With tech giants like Meta offering more competitive pay, Apple risks losing key talent. Financial considerations are also at play: Anthropic is reportedly demanding a multi-billion dollar annual licensing fee, pushing Apple to weigh OpenAI as a more flexible alternative if negotiations stall.

Should Apple move forward with a third-party AI partnership, it could marginalize AI chief John Giannandrea, a staunch advocate for in-house model development. His influence is already waning, with many AI teams now reporting directly to Federighi. Projects like “Swift Assist” have been shelved in favor of integrating external tools like ChatGPT. While Apple continues to develop its own models for select use cases, its willingness to partner externally highlights a major strategic pivot in its AI ambitions.

Source: The Sun