An analyst's insight suggests Apple's strategy to license Google's Gemini Large Language Model (LLM) while simultaneously developing proprietary, privacy-centric local AI models could position Apple advantageously in the AI landscape. As Apple gears up to amplify its AI offerings in 2024, this dual approach seems to draw from the company's historical playbook of leveraging its strengths.
A JP Morgan analysis highlighted by AppleInsider indicates that Apple's potential collaboration with an LLM provider like Google, coupled with its efforts to create compact local models, could set it apart from competitors. This strategy aligns with Apple's emphasis on on-device processing capabilities, prioritizing user privacy over the construction of a standalone, extensive LLM.
Reports have surfaced about Apple negotiating with Google to integrate its Gemini LLM into the iPhone ecosystem. This development coincided with Apple's disclosure of research on MM1, a smaller, locally operable pre-trained model. These movements suggest a complementary strategy where Apple could offer the expansive intelligence of Gemini while ensuring user privacy through localized MM1 models.
This approach mirrors Apple's existing relationship with Google regarding search functionalities, where Google serves as the default web search engine, and Apple enhances user privacy through its local search tool, Spotlight. Further leaks about an internal tool named Ask, designed to utilize a local knowledge base and touted for its adaptability and database flexibility, reinforce Apple's strategic direction.
JP Morgan's analysis also touches on the economic implications of this strategy, pointing to potential savings on infrastructure and enhanced user experiences in applications. With an "overweight" rating and a price target of $215 for Apple, JP Morgan's outlook underscores the financial viability and consumer appeal of this dual-path AI strategy.
