×
×
homepage logo
SUBSCRIBE

Tech Matters: Apple’s On-Device AI — A safer, greener approach?

By Leslie Meredith - Special to the Standard-Examiner | Sep 18, 2024

Photo supplied

Leslie Meredith

It’s that time of year again: school is back in full swing, Starbucks returns Pumpkin Spice Lattes to its menu and Apple launches its latest iPhones. There were no big surprises at Apple’s annual event. The new iPhone 16 lineup includes the usual camera upgrades, the pricing is the same as last year’s models and all four iPhones have been upgraded with the new A18 chip, designed to bring artificial intelligence processing power to the phones. We won’t see much of its branded AI called Apple Intelligence for at least a month or two.

A few simple AI features will be released with iOS 18.1 in October with the big ones like Chat GPT integration to follow at an unspecified date. But there’s a bigger story here that will unfold that involves where this computing takes place. The fact that Apple has designed its new chip to process the majority of AI functions on the device itself rather than at remote data centers marks a significant difference between the way AI processing has traditionally been done and how Apple plans to do it. What does that mean for you? Big privacy gains and potential energy savings.

Currently, AI models, particularly large language models, require vast amounts of computational power and energy to train and run. For example, a single ChatGPT query uses nearly 10 times more electricity than a Google search, according to a Goldman Sachs Research report. The International Energy Agency estimates data center electricity consumption in 2026 will be double that of 2022, reaching about 1,000 terawatts.

In contrast, Apple’s solution to keep as much of the AI workload local as possible, reduces the need for data to travel back and forth between servers and devices. This cuts down on energy consumption tied to data transmission and network use. As a result, everyday AI tasks — like text and image generation — can be done without adding to the energy demands of a cloud network.

There are also differences in the chips themselves. For instance, Nvidia’s latest AI chips consume up to 300% more power than previous generations. Apple’s custom silicon, like the Neural Engine in its new A18 chips, is specifically designed for efficient AI processing. This allows for better performance with lower power draw compared to general-purpose processors.

But higher intensity tasks do require more computing power than what’s available on an iPhone or other Apple device. For these tasks, Apple has introduced Private Cloud Compute, a server-based model that runs on dedicated Apple silicon servers. This hybrid model gives devices the ability to shift computational capacity between the device and servers at an Apple data center as a way to optimize computing power and minimize energy use.

You might wonder about the impact of this additional processing on the life of the battery. Because of the new chip design coupled with AI processing on demand, meaning that AI processing does not run in the background but springs into action only when needed for tasks like language translation and augmented reality applications, battery life may be unaffected — we’ll have to wait and see when the more robust AI features become available.

Energy savings are not the only benefit to on-device AI processing. Privacy is part of the package too. By keeping as many AI computations on the device as possible, Apple minimizes the amount of personal data that needs to be sent to the cloud, reducing potential privacy risks. The company has also taken a privacy-first approach with its servers.

For instance, imagine using Siri to ask a question or using AI-powered photo sorting in the Photos app. In a cloud-based system, your voice recording or photos would typically be sent to remote servers for analysis. With Apple Intelligence, much of that processing will happen on your iPhone. This shift not only reduces the risk of sensitive data being intercepted but also cuts down on the amount of data stored on remote servers — and the possibility of a data breach.

While the most noteworthy AI tools won’t be released in the immediate future, Apple did reveal what it is developing in June. We can look forward to “Writing Tools,” “Priority Messages” and “Quick Reply” in Mail; “Priority Notifications” on your lock screen; “Reduce Interruptions” mode for Focus; the ability to ​​record, transcribe and summarize audio in both Notes and Phone apps; and Image Playground where you can quickly generate images in three styles: animation, illustration and sketch.

The new iPhone 16 series goes on sale at the end of this week. If you’re ready for a new phone, this one is worth consideration. You will be future-proofing your mobile computing and contributing to much-needed energy conservation.

Leslie Meredith has been writing about technology for more than a decade. As a mom of four, value, usefulness, and online safety take priority. Have a question? Email Leslie at asklesliemeredith@gmail.com.