Apple Taps Google’s Gemini Models to Power Siri AI Beta
Apple integrated Google's Gemini Flash-Lite, Flash, Pro and Image AI models to power its Siri AI beta launching later this year, while maintaining its own proprietary Apple Frontier Models trained via reinforcement learning. Craig Federighi clarified that Apple does not use Google Search or Google’s customer-deployed model infrastructure, preserving data privacy and intellectual property separation.
1. Partnership Overview
At its Worldwide Developers Conference, Apple unveiled Siri AI, a redesigned assistant set to enter beta later this year. The update leverages Google’s Gemini Flash-Lite, Flash, Pro and Image models to deliver advanced language understanding and multimodal capabilities within Siri.
2. Model Use and Infrastructure Separation
Craig Federighi explained that Apple built its own Apple Frontier Models using proprietary data and reinforcement learning, then refined them with outputs from Google’s Gemini Frontier models. Apple does not rely on Google Search or Google’s customer-deployed infrastructure to ground its assistant in world knowledge.
3. Strategic Implications
This licensing deal positions Google as a key AI provider beyond search, potentially boosting revenue for Google Cloud AI services. It also strengthens Google’s footprint in consumer devices, highlighting the value of its Gemini models and reinforcing GOOGL’s growth outlook through diversified AI partnerships.
