Apple launched several AI-powered features under the Apple Intelligence system at last week’s Worldwide Developers Conference (WWDC). These features will be coming in late 2024 and the first half of 2025 – marking a shift in rollout strategy for the company.
According to Bloomberg’s Mark Gurman, Apple not following a strict timeline when it comes to rolling out AI features to users has several benefits for the Cupertino-based tech giant. Gurman, however, noted that Apple Intelligence’s initial rollout with iOS 18 will include several functionalities such as leveraging AI to prioritise notifications, provide summaries of alerts and text messages as well as new writing tools, image generation and custom AI-generated emoji (Genmoji)..
These capabilities are said to come at a later stage, and won’t be available in the first beta version of iOS 18. Apple typically releases the new iOS version publicly in September, coinciding with iPhone sales launches. Since iPhones require manufacturing and operating system installation weeks in advance, features need to be finalised by mid-August.
This time around a portion of the AI services will be delayed until next year due to these time constraints. Gurman has also listed out a bunch of reasons why this slow-roll of AI feature is the right choice for the company.
Dodging the staffing constraints
A phased rollout of AI features will help Apple prevent staffing bottlenecks, allowing engineers to focus on specific features, release them and then move on to others. This strategy will help the company avoid overwhelming developers and past issues with rushed launches.
Language prioritisation
Launching AI features with American English first allows more time to train AI models for other languages – which is a complex task. Global availability will take time but this will also help the company to better test its, which in turn provide better functionality from the beginning.
Scalability and stability
Recently, a report said that Apple is training its model on Google’s TPUs as it does not have a large and sophisticated infrastructure like Google or Microsoft. Apple is building its cloud infrastructure to handle Apple Intelligence features. A gradual rollout reduces the risk of network crashes from a sudden surge in users. This approach also benefits OpenAI, which partners with Apple and utilises Microsoft servers for its chatbot.
Mitigating misinformation
One of the main problems with AI-generated responses is misinformation. Starting with a smaller user base will allow Apple for better monitoring and quicker fixes for potential AI-generated misinformation (like OpenAI’s ChatGPT). This limits the spread of inaccurate information.
According to Bloomberg’s Mark Gurman, Apple not following a strict timeline when it comes to rolling out AI features to users has several benefits for the Cupertino-based tech giant. Gurman, however, noted that Apple Intelligence’s initial rollout with iOS 18 will include several functionalities such as leveraging AI to prioritise notifications, provide summaries of alerts and text messages as well as new writing tools, image generation and custom AI-generated emoji (Genmoji)..
These capabilities are said to come at a later stage, and won’t be available in the first beta version of iOS 18. Apple typically releases the new iOS version publicly in September, coinciding with iPhone sales launches. Since iPhones require manufacturing and operating system installation weeks in advance, features need to be finalised by mid-August.
This time around a portion of the AI services will be delayed until next year due to these time constraints. Gurman has also listed out a bunch of reasons why this slow-roll of AI feature is the right choice for the company.
Dodging the staffing constraints
A phased rollout of AI features will help Apple prevent staffing bottlenecks, allowing engineers to focus on specific features, release them and then move on to others. This strategy will help the company avoid overwhelming developers and past issues with rushed launches.
Language prioritisation
Launching AI features with American English first allows more time to train AI models for other languages – which is a complex task. Global availability will take time but this will also help the company to better test its, which in turn provide better functionality from the beginning.
Scalability and stability
Recently, a report said that Apple is training its model on Google’s TPUs as it does not have a large and sophisticated infrastructure like Google or Microsoft. Apple is building its cloud infrastructure to handle Apple Intelligence features. A gradual rollout reduces the risk of network crashes from a sudden surge in users. This approach also benefits OpenAI, which partners with Apple and utilises Microsoft servers for its chatbot.
Mitigating misinformation
One of the main problems with AI-generated responses is misinformation. Starting with a smaller user base will allow Apple for better monitoring and quicker fixes for potential AI-generated misinformation (like OpenAI’s ChatGPT). This limits the spread of inaccurate information.
