It's more difficult to describe the gear because it varies depending on the manufacturer.
Most analysts who talked with CNBC agreed that new gadgets will have more sophisticated
chips to run artificial intelligence (AI) programs and that those apps will operate locally on
the device rather than in the cloud.
Chipsets for smartphones have been introduced by companies like Qualcomm and MediaTek,
enabling the computing power needed for AI applications.
However, AI technology in phones is not new. Some AI-powered functions, including
background blur effects on smartphones and picture editing have been present in devices
for years.
Large language models and generative AI are fresh developments. Huge AI models called
"large language models," which are trained on enormous volumes of data, form the
foundation for products like the commonly used chatbots. These models open up
additional functionality, like chatbots' capacity to produce text or graphics in response to
user input.
"Having a chatbot is not the only thing; we have had these virtual assistants for a long
now. The distinction is that it is now generative, enabling them to compose a poem or
provide a meeting summary. When it came to creating images from words, that was a first,
according to Ma.
The phrase "on-device AI" is another important piece of the smartphone AI puzzle. Previously,
a lot of AI apps were downloaded onto phones after being partially processed in the cloud.
However, more AI applications will probably be performed entirely in the device rather than
in a data center due to improved CPUs and the possibility for massive language models to
effectively become smaller. According to smartphone manufacturers, on-device AI increases device security, opens up
new applications, and speeds things up because processing is done on the device itself.
According to Ma and Wood, this might allow developers to construct new applications.
"Anticipatory computing," or the notion that artificial intelligence (AI) "is smart enough to
learn your behavior as a user make the device so much more intuitive and predict what
you want to do next without you having to do much," is what smartphone makers ultimately
hope to accomplish, according to Wood.
However, are AI phones already a thing?
Kind of. While AI has long been included in gadgets, the latest wave of on-device AI with
massive language models is still in its infancy.
Many AI-powered features will be on display by device manufacturers at MWC, and some of
them are currently available. Samsung touted its AI capabilities while launching its flagship
Galaxy S24 smartphone line in January. The ability to circle an image or text on any app and
instantly search it on Google was one feature that caught people's attention.
AI feature demonstrations, ranging from smartphone chatbots to camera apps, are probably
going to take place at MWC.
However, as IDC's Ma points out, many of these benefits really rely on cloud processing and
are not truly on-device. It will take a "number of years" before third-party developers find a
"killer use case or that compelling use case that consumer can't do without," he continued,
even with AI capabilities on smartphones.
According to Wood, the risk lies in the fact that producers of smartphones tend to focus more
on AI than on the experiences that their products can provide for customers.
"We are building an unbelievable foundational platform for AI on devices. 2024 will be the year
we look back on and say that's where it all started to happen but it could be a long time before
we start of these benefits in terms of game-changing experiences," Wood said. "Consumers
have no idea what an AI smartphone is, they need the use cases to go around it." The risk is
that there will be "AI fatigue." In the end, the grandiose AI experiences smartphone makers
are dreaming of may be a long way off.
Comments
Post a Comment