SiriSummarization private framework within the code has been making calls to OpenAI's ChatGPT API. This integration plays a pivotal role in Apple's internal testing of evolving AI capabilities, as evident from system prompts like "please summarize," "please answer this question," and "please summarize the given text," as reported by 9to5Mac. Moreover, these system prompts suggest an integration of AI features within the Messages app, aligning with earlier reports from Bloomberg that hinted at Apple's efforts to integrate a Message app capable of handling queries and auto-completing sentences.
The report adds that Apple's collaboration with ChatGPT is not about dependence but rather about setting benchmarks. The SiriSummarization framework primarily utilizes on-device models for summarization, showcasing Apple's dedication to leveraging its proprietary AI models. The iOS 17.4 code reveals active testing of four distinct AI models, including the internally developed "Ajax." Notably, there are two versions of AjaxGPT—one processed on-device and the other off-device.
What stands out in this revelation is Apple's dual approach to AI integration. While actively developing its own AI system, including models like "Ajax," the company is simultaneously benchmarking against external models like ChatGPT and FLAN-T5. This strategic move aligns with Bloomberg's October report by Mark Gurman, shedding light on Apple's AI goals for iOS 18.
The report outlined Apple's directive to infuse substantial language model capabilities into the upcoming iOS, promising users a transformative AI-driven experience. As iOS 18 approaches its release, Apple's intent to advance AI capabilities. Milestone Alert!
Livemint tops charts as the fastest