New news: Apple feels Meta has a bad reputation on privacy, so it won’t integrate its AI model
Brook.X
Recent news indicated that Apple had been in discussions with companies like Meta and Google to integrate their AI models into Apple Intelligence. This move would allow users to choose from various AI service providers, not limited to OpenAI alone.
Apple aimed to collaborate with top AI industry players such as Anthropic and Perplexity, where such a partnership wouldn't involve any payment from Apple to these AI firms. Instead, these companies would gain access to a broader user base through the Apple ecosystem and attract users to subscribe to premium versions.
However, the latest information reveals that Apple probably never seriously considered partnering with Meta. The reason is straightforward: despite the Meta Llama series models being exceptional, Meta's reputation in privacy matters is notoriously poor.
As a social media giant, Meta's primary business has been selling user data to advertisers for targeted advertising. Over the past decade, Meta (formerly known as Facebook) has faced numerous data breaches and allegations of improperly selling user data to advertising clients.
Apple, known for its stance on protecting user privacy, likely fears that collaborating with Meta could tarnish its privacy reputation. Therefore, it prefers not to engage in a partnership with Meta.
Other companies, including Google, though reliant on advertising, have taken far better measures regarding privacy protection. Hence, users might see Google's Gemini on the iPhone, but Meta's Llama model seems unlikely to make the cut.
Meta has not yet responded to these rumors. The Meta AI Research Institute's Llama series models are indeed exceptional in capability. These models are open but not open-source. Large companies looking to use the Llama model must obtain a license from Meta, while small to medium AI firms can directly use and adjust the Llama model for training their AI systems without a commercial license.
