Apple has replaced its top artificial intelligence executive following significant product failures and internal dissent, signaling a potential shift in its strategy amid an ironic reliance on rival Google’s technology.
On December 1, 2025, Apple announced that John Giannandrea, who had led the company’s AI efforts since 2018, would step down from his role. He will continue as an advisor until next year.
Taking over the crucial AI leadership position is Amar Subramanya, a former senior executive at Microsoft with a notable background in engineering for Google’s Gemini Assistant. His past experience with Apple’s longtime rival is particularly significant.
Giannandrea’s departure follows accumulating criticism of Apple Intelligence, which launched in October 2024. The product, rather than being a breakthrough, reportedly caused user frustration and garnered significant negative feedback.
Among the specific issues were reports of the notification summarization feature generating inaccurate or false information. The BBC notably complained twice about Apple’s AI misrepresenting news.
These incidents reportedly damaged Apple’s reputation for precision and reliability.
A May 2025 Bloomberg report detailed further internal struggles. Craig Federighi, head of software, reportedly discovered that many features of a new Siri version were non-functional during his own testing.
This led to an indefinite delay in Siri’s public rollout. Subsequently, the company faced class-action lawsuits from iPhone 16 purchasers who felt misled.
Internal working conditions also reportedly deteriorated. Some employees were said to have mocked Giannandrea’s team as “AI/MLess,” a term reflecting perceived failures in artificial intelligence and machine learning.
The period also saw a “brain drain” as talented AI researchers left Apple for competitors such as OpenAI, Google, and Meta.
The current situation has compelled Apple to take an unprecedented step: relying on Google Gemini to power upcoming versions of Siri. This marks an ironic development for two companies that have been intense rivals across technology sectors for over 15 years.
Subramanya’s primary challenge will be to reconcile Apple’s foundational emphasis on user privacy with the demand for advanced AI capabilities. Apple has steadfastly committed to on-device processing via its Apple Silicon chips to prevent data leaks.
This approach contrasts sharply with competitors who invest heavily in massive data centers to process large AI models. Apple’s refusal to collect extensive user data, while upholding privacy, limits the raw material available for training its AI.