Apple has appointed Amar Subramanya, a veteran of both Google and Microsoft, as its new head of Artificial Intelligence. The move comes as John Giannandrea, who led Apple’s AI efforts since 2018, steps down after a period marked by significant product missteps and internal turmoil.
The Shift in Leadership
Subramanya’s arrival signals a clear attempt by Apple to regain ground in the rapidly evolving AI landscape. He brings extensive experience from leading engineering roles at Microsoft and, crucially, Google, where he recently oversaw development of the Gemini Assistant. This background is seen as particularly valuable, given Subramanya’s deep understanding of Apple’s primary competitors.
Apple Intelligence Struggles
The change at the top follows a series of high-profile failures with Apple Intelligence, the company’s AI-powered features introduced in late 2024. Early versions of the platform produced demonstrably false information, including incorrect reporting on high-profile cases like the Brian Thompson murder investigation and premature championship claims in professional darts.
These inaccuracies led to public embarrassment, complaints from news organizations like the BBC, and ultimately, legal challenges. A delayed launch of the upgraded Siri, which was supposed to be a centerpiece of the Apple Intelligence rollout, triggered class-action lawsuits from consumers who had pre-ordered devices based on promised AI capabilities.
Internal Dysfunction
According to a Bloomberg investigation, Apple’s AI division suffered from organizational dysfunction, weak communication between teams, and budget misalignments. The report revealed that Giannandrea had been sidelined months before his departure, with key responsibilities like Siri development stripped from his control. Employees reportedly mocked the AI division internally as “AI/MLess,” while many researchers left for competitors like OpenAI, Google, and Meta.
Reliance on Google’s Gemini
In a striking reversal, Apple is now reportedly relying on Google’s Gemini to power the next iteration of Siri. This marks a major shift for the company, given its historical rivalry with Google across numerous tech sectors. The move underscores Apple’s need to catch up quickly in AI, even if it means partnering with a long-time competitor.
Apple’s Unique Approach
Apple has long emphasized a privacy-first approach to AI, processing tasks on-device using its custom Apple Silicon chips to avoid collecting user data. While this strategy has resonated with privacy advocates, it has also resulted in less capable AI models compared to competitors who leverage massive data centers and real-world user information for training.
The success of Apple’s approach remains uncertain. On-device models are inherently limited by computational power and data availability, while the company’s reluctance to collect user data has hampered its ability to train competitive AI systems.
The question now is whether Apple can overcome these trade-offs or if its privacy-focused strategy will leave it permanently behind in the AI race. The appointment of Subramanya represents a gamble that his expertise can bridge the gap, but the underlying challenges remain significant.
