Apple has announced a significant shift in its approach to building Siri features, allowing engineers to use third-party Large Language Models (LLMs) from external companies for the first time under Craig Federighi’s leadership. This move comes after internal conflicts and restrictions on using open-source models in Apple’s software products.
Prior to Federighi’s tenure, Apple engineers were only allowed to test their own LLMs against third-party models during development. However, they were not permitted to use these external models in the final product. The responsibility of building these models previously fell to Giannandrea’s team, which found that Apple’s models did not perform as well as OpenAI’s technology.
Under Federighi’s leadership, the company has now relaxed its restrictions on using open-source LLMs. This move is expected to benefit users who prioritize high-quality Siri features over the underlying technology. The shift in approach may also lead to improved performance and functionality for Siri, making it a more competitive virtual assistant.
The new policy allows Apple engineers to explore various third-party LLMs, including those from OpenAI, to develop AI-powered features that meet user expectations. This move is seen as a significant step forward for Siri’s development, and its impact on the company’s future plans remains to be seen.
Source: https://9to5mac.com/2025/04/10/craig-federighis-leadership-has-already-resulted-in-this-major-siri-pivot-per-report