Big Tech’s Mobile Assistants Are Still Hungry for Your Data

Key Takeaways

• Mobile platforms like Apple Intelligence and Galaxy AI combine cloud-based and on-device artificial intelligence.
• This approach aims to provide personalized assistants without sacrificing privacy.
• However, Big Tech providers still benefit from controlling access to user data.

New Mobile AI assistants that have emerged in 2024 promise to deliver personalized AI experiences without sacrificing privacy.

But just because platforms like Apple Intelligence diverge from the cloud-centric model used by ChatGPT doesn’t mean Big Tech mobile players don’t still benefit from user data. 

How Mobile AI Uses Personal Data

Mobile AI combines on-device models and cloud AI services to personalize agents for each user.

This multi-model approach has created privacy challenges for device manufacturers, who can be reluctant to outsource data processing to third parties like OpenAI.

To overcome this challenge, some of the space’s leading players have redesigned their data management systems for AI. 

For example, Apple introduced “Private Cloud Compute,” a new AI delivery model that only sends the minimum necessary data to Apple’s servers for processing without the company retaining unnecessary device information. 

Meanwhile, Samsung’s Galaxy AI provides settings that let users switch between on-device and cloud processing, allowing them to avoid the cloud entirely if they want. 

However, although major mobile players have introduced solutions that minimize how much data they can harvest through cloud services, they still have an incentive to control the flow of user data collected by AI assistants. 

Interoperability Challenges

While on-device or edge AI solves some of the privacy challenges associated with the cloud, it creates a new problem for users. Or rather, a new version of an old problem—friction between technology ecosystems. 

To fully realize the potential of personalized AI, agents must be able to draw data from users’ devices. However, smartphone manufacturers, especially Apple, have rarely supported interoperability with rival brands. 

Apple Intelligence Controls Data Access

“I see the way Apple is building [mobile AI] is basically replicating their success with iOS and the App Store,” PIN AI Chief Scientist Bill Sun told CCN.

“I can imagine very easily that they will still charge the 30% Apple tax for all the in-app transactions they can. They will still basically ask everyone to develop in their ecosystem.”

As Sun explained, while Apple Intelligence lets iPhone users retain their data on their devices, Apple still holds the keys developers need to access this data. 

Going forward, as mobile assistant adoption rises and more apps start to integrate AI, these keys could become the next major barrier to device interoperability, erecting yet another wall between already segregated ecosystems. 

Meanwhile, Sun speculates that Apple could move to monetize its position as device manufacturers, model developers, and other stakeholders figure out how to generate revenue from the new technology.