Apple faces criticism over the exclusivity of artificial intelligence in new devices

Apple has announced a significant step forward in integrating artificial intelligence (AI) across its product lineup. This new initiative, known as Apple Intelligence, promises to revolutionize the way users use iPhones, Macs and iPads. Apple Intelligence, expected to launch later this year, aims to empower devices with smarter and more intuitive features, though it offers a certain exclusivity that may make some users feel left out.

The next frontier in AI

Apple Intelligence isn’t just about making Siri smarter or generating creative content like emails, essays, and emojis. It represents a vision of the future where artificial intelligence integrates seamlessly with everyday tasks, providing a user-friendly and efficient experience. While not entirely groundbreaking, Apple’s approach aims to deliver one of the most accessible and useful implementations of generative AI to date.

But there’s a catch: only certain Apple devices support Apple Intelligence. In particular, it requires iPhone 15 Pro or iPhone 15 Pro Max. Regular iPhone 15 users won’t be able to take advantage of these new AI features, making their relatively new phones partially obsolete in this regard. Mac users, on the other hand, will need an Apple Silicon computer, the line of which will launch in 2020.

Exclusion and cynicism

The requirement for the latest equipment has sparked some criticism. Some see this as a ploy to ensure frequent updates, in line with typical phone and laptop update cycles. While it may be common to update your phone every year, it’s less common for a laptop and often seen as unnecessary.

The technical reality is more nuanced. The computing requirements of Apple Intelligence are very different from standard tasks performed on an iPhone or Mac. Most generative AI work these days relies on remote cloud servers that do the heavy lifting, with the final results pushed back to the user’s device. This method, similar to digital assistants like Siri or Alexa, is efficient but limits real-time processing on the device.

Apple aims to change that by emphasizing artificial intelligence processing on devices, a move that follows the company’s long-standing commitment to privacy.

Privacy and on-device processing

Craig Federighi, Apple’s senior vice president of software engineering, emphasized privacy as a key driver of on-device computing. “You shouldn’t hand over all the details of your life to be stored and analyzed in someone else’s AI cloud,” Federighi said during the announcement. Apple has deeply integrated this processing into its devices and apps, ensuring that personal data remains on the device rather than being sent to external servers.

This approach reflects Microsoft’s strategy with the Copilot+ standard, which also emphasizes AI processing on devices, but primarily in the case of laptops. Both Apple and Microsoft rely heavily on artificial intelligence developed by OpenAI, the creator of ChatGPT.

NPU: the heart of AI processing

The centerpiece of this new wave of on-device AI is the Neural Processing Unit (NPU). This specialized chip is designed to handle multiple operations simultaneously while minimizing power consumption. Since 2017, Apple has introduced NPUs, called Apple Neural Engine (ANE), in iPhone 8, iPhone 8 Plus and iPhone X.

The role of the NPU is crucial in managing the complex computations required for AI tasks. These tasks include matrix manipulation, the core process behind chatbots and other generative artificial intelligence. Despite advances in NPU technology, there are still significant differences in capabilities between consumer devices and the dedicated AI hardware used in cloud servers.

The reality of hardware limitations

Despite the capabilities of modern NPUs, consumer devices like the iPhone 15 Pro Max still face limitations. The NPU in these devices, while powerful, cannot match the vast resources of cloud-based AI systems. For example, the latest Nvidia H200 graphics cards used in AI servers offer bandwidth of 4.8 TB per second, a level that far exceeds the capabilities of any consumer-grade device.

Moreover, memory constraints play a crucial role. Older iPhones, despite having powerful NPUs, lack enough RAM to locally support advanced AI models. This is the likely reason why Apple Intelligence requires the latest iPhones with 8GB RAM, as running AI models locally requires significant memory resources.

The future of Apple intelligence

As Apple prepares to implement Apple Intelligence in iOS 18, the true extent of its capabilities and limitations will become clearer. While on-device AI processing provides greater privacy and efficiency, the actual use of these features may vary depending on your device’s hardware.

Users are wondering whether future versions of the iPhone, such as the iPhone 16, will offer the full suite of Apple Intelligence features, or whether Apple will continue to reserve these features for its premium Pro series devices. We’ll find out the answer with the public release of iOS 18, which is expected to launch in September alongside iPhone 16.

Apple Intelligence is a significant step in the integration of artificial intelligence into consumer technologies, promising smarter and more intuitive devices. But its true impact will depend on how well Apple balances innovation with accessibility, ensuring that the benefits of AI are available to a wider range of users, not just those equipped with the latest hardware.