Here comes the AI phone

Smart devices have been the biggest single driver of mobile revenue for the past 15 years, so it makes sense to try to understand the new generation of AI phones.

Analyst firm Canalys has forecast that AI-capable devices will account for up to around 5% of smartphones shipped in 2024 – that still means around 60 million units. Canalys believes the smartphone will play a central, if not the central, role in the era of on-device AI, thanks to its huge installed base, its versatility and its portability.

In a new post, analyst Lucas Zhong says AI has already been running on devices for nearly a decade.

“Chipset vendors such as Qualcomm, Mediatek, Samsung and Google have focused on improving NPU [neural processing unit]/TPU [tensor processing unit] performance for years. Smartphone vendors including Apple, Huawei, vivo, and Xiaomi are implementing AIML algorithms locally to improve imaging quality, battery life, and typing experience,” he writes.

Automated social media posts

But with the arrival of large language models (LLMs) and other GenAI models, on-device AI will require a different computing platform and new software capabilities.

To get a sense of what the new phones will be able to do, Canalys has come up with a hierarchy of likely use cases, running from low-frequency professional apps like scheduling to high-frequency background tasks like battery optimization (see graphic above).

These might include cross-device automation, like harnessing sensors to automatically turn on the lights or TV when the user arrives home; automating social media posts and photo editing; or delivering automated call answering.

To read the complete article, visit Light Reading.