Apple is entering the realm of generative artificial intelligence in collaboration with the unexpected partner, OpenAI.
During the Worldwide Developer Conference in Cupertino, California, Apple’s CEO, Tim Cook, unveiled the highly anticipated AI transformation, referred to as “Apple Intelligence.” This transformative initiative encompasses features that will revolutionize the iOS ecosystem in both significant and subtle ways. Notably, Siri, Apple’s current voice assistant, has undergone a profound overhaul utilizing generative AI technology.
Moreover, Apple disclosed its plan to integrate external AI models into its software, commencing with OpenAI’s ChatGPT later this year. Importantly, this integration will be voluntary and will not mandate a ChatGPT subscription. Siri will assess if a given query could benefit from leveraging ChatGPT and seek permission from users before sharing data with the model. Apple also hinted at leveraging other AI models in the future.
Historically, Apple has not been prominently featured in conversations surrounding generative AI, in stark contrast to competitors like Google and Microsoft who swiftly adopted the technology following the emergence of ChatGPT in late 2022. Apple seems to have meticulously crafted a comprehensive strategy that encompasses both on-device AI processing and cloud-based AI solutions.
Craig Federighi, Apple’s senior vice president of software engineering, emphasized the company’s commitment to privacy and security within its new AI paradigm. Alongside advocating for privacy-conscious on-device AI applications, Federighi introduced “Private Cloud Compute,” a technology designed to safeguard data during more resource-intensive cloud-based AI operations. He assured users that their data remains secure and inaccessible to Apple.
Apple showcased the capabilities of Apple Intelligence across various applications, from rewriting emails in Mail to generating custom emojis on demand and summarizing content in Safari. The revamped Siri, powered by AI enhancements, promises improved comprehension of complex voice commands and better information retrieval even in situations with ambiguous queries.
Additionally, Apple will introduce compact generative AI algorithms that operate locally on devices, aiming to enhance responsiveness and fortify user data protection. Striking a delicate balance between prioritizing user privacy and security while embracing generative AI advancements presents a significant challenge for the company.
Speculations preceding WWDC suggested an impending partnership between Apple and OpenAI. Despite the apparent disparity in their approaches, Apple’s cautious product development ethos, user-centric focus, and aesthetic excellence complement OpenAI’s drive for innovation and experimentation.
To mitigate risks, Apple primarily relies on its in-house AI solutions for integrations, while ensuring that the adoption of ChatGPT remains a step-by-step opt-in process. Though showcasing impressive demonstrations, the real-world performance of Apple Intelligence remains unverified.
“With Apple’s latest announcement, major tech players have outlined their strategies for generative AI,” says Any Wu, an associate professor at Harvard Business School specializing in AI applications in tech. Wu notes that despite entering the field later, Apple’s expansive user base presents a significant advantage.
Nonetheless, Wu highlights the challenges of managing the costs and risks associated with deploying potent generative AI models, especially in combating misleading information. He observes a potential conflict with Apple’s ethos of delivering meticulously crafted products under complete control.
This narrative is continuously evolving. Stay tuned for updates or follow our live coverage of WWDC here.