[Into the World of AI] Apple’s AI Emphasizes Privacy Protection: “No User Data Used for Training”
Apple, once labeled as an "AI underachiever," has launched a counterattack by unveiling a host of new AI features. While the company built its own AI model based on Google's AI model Gemini—seemingly acknowledging the limitations of developing a proprietary model from scratch—Apple placed particular emphasis on protecting user privacy.
Craig Federighi, Senior Vice President of Software Engineering at Apple, highlighted this focus during the annual developer event, World Wide Developers Conference (WWDC) 2026, held at Apple headquarters in Cupertino, California, on June 8 (local time). He stated, "Most AI vendors shift the responsibility of data protection to the user, but Apple ensures that neither Apple nor anyone else can access or store user data by utilizing on-device AI and Private Cloud. We've blocked such access entirely."
On the 8th (local time), attendees at the Worldwide Developers Conference (WWDC) held at Apple's headquarters 'Apple Park' in Cupertino, California, captured images of Apple CEO Tim Cook. Photo by Yonhap News
View original imageApple introduced "Siri AI," which integrates AI capabilities into the voice assistant Siri using Google's large language model (LLM), Gemini, as a foundation. Siri AI can understand user context and partially takes on the role of an "agentic AI" capable of performing tasks on the user's behalf. It can recognize previous conversations, photos, emails, and other contextual information, as well as add events to the calendar or items to the reminders list.
Throughout the unveiling of Siri AI, Apple placed special emphasis on privacy protection. Federighi stressed, "Even after integrating the Gemini model, user information will not flow to Google," adding, "We do not use Google’s general consumer model."
Apple first built the core framework of the Apple Foundation Model (AFM), which powers Siri AI, and then refined it by leveraging Gemini. Since the company does not use the Gemini model directly, user input data is not transferred to Google, according to Federighi. While Siri AI utilizes user information stored on devices such as iPhones and MacBooks during operation, this process is thoroughly protected to ensure that personal data is not leaked.
Security is prioritized at every stage where data interacts with the AI model. Apple applies an "on-device" approach by embedding a lightweight AI model within the device, enabling simple requests to be processed locally. In this way, data is handled solely on the device, minimizing the risk of information leakage.
Requests that require more complex processing are handled by "AFM Cloud Pro," a high-performance model running on Apple’s servers. In this process, the content of user requests and data are transmitted to Apple’s servers through a Private Cloud Compute (PCC) environment.
Hot Picks Today
"Two Spend 500,000 Won in a Flash, Yet Wait Two Hours?" ... 'Dreamland' Shunned by Kids and Foreign Tourists
- The Era of Getting 200 Million Won After 30 Years Is Over... A Japanese Bank's "Perpetual Department Head" Who Just Clocked In Cries Out
- Spain Hit by Its Worst Tragedy... Lee Kangin, Moreno and Real Madrid Join Soccer World in Mourning
- They Said, “Make It Your Home for Life” - Then Suddenly Kicked 200 Passengers Off... What Happened to the Around-the-World Cruise?
- "130 Rice Bowls Sold Out in a Day"... 'Unfiltered Face' Adds 8 Million Won to Night Market Sales: China's Bungeunwoo
Apple has consistently implemented policies that prioritize user privacy. In 2021, Apple introduced the App Tracking Transparency (ATT) policy, requiring platform companies such as Google and Meta to obtain user consent before utilizing app usage information for advertising purposes. As a result, these platform companies could no longer run personalized advertisements as before, leading to a decline in their revenue.
© The Asia Business Daily. All rights reserved. Unauthorized AI training and use prohibited.