During the CES 2024 keynote speech at the Intel Client Open House at the Venetian Hotel and Resort, Las Vegas, Intel’s Client Computing Group head, Michelle Johnston Holthaus, stressed her company’s focus on AI technology. Joined on stage by partners from Dell, HP, Lenovo, and Microsoft, the discussion highlighted the impact that onboard AI technology, namely the neural processing unit, will have on the consumer user experience.
The AI technology featured in Intel’s new Core Ultra 7 processors will allow users to employ AI without an Internet connection or the need to send personal data into the cloud. Notebooks with Intel Core Ultra 7 processors can utilise the dedicated NPU for AI operations, from image generation to audio isolation, leaving the CPU and GPU free for other tasks.
As well as showcasing several notebooks featuring the Core Ultra 7 mobile processors, the new MSI Claw portable gaming console was also revealed. With a similar form factor to the Steamdeck and ROG Ally, the MSI Claw is a Windows-based gaming platform. It is powered by the Intel Core Ultra 7 and the onboard Intel ARC GPU. This is an interesting development as, for the first time, it places Intel’s ARC graphics technology in a dedicated gaming platform.
With AI advancements being a popular concept in the public eye, it is easy to dismiss Intel’s AI focus as a marketing ploy seeking to exploit the current zeitgeist. The inclusion of an NPU (effectively a parallel computing processor similar to a GPU), joining the CPU and GPU in new the Intel Core Ultra processors, is a real step forward, away from the more nebulous concept of cloud-based AI.
Looking to the future, Johnston Holthaus revealed that 2024 would see Intel release its Arrow Lake processors, the first desktop processors with AI accelerators onboard. As well as Arrow Lake chips, 2024 will also see the release of Lunar Lake AI-equipped processors for portable devices.
It is still the early days of AI technology, and neural processing units are at a similar point to that of graphic processing units/3D accelerators twenty-five years ago. Applications need to support the hardware to properly take advantage of it. According to Intel, software vendors are very keen to support this technology, and I don’t doubt this. Adobe Photoshop, for instance, currently uses cloud-based AI technology that could easily be switched to use a device’s onboard AI capabilities. I’ve seen the open-source image editor, GIMP, using the Core Ultra 7’s NPU with Stable Diffusion, so it’s only a matter of time before using this technology is routine.
AI evangelism aside, for the consumer, the benefits of this AI revolution may not be immediately apparent. Up until now, access to AI computing has been walled off behind subscriptions and server-based AI models. Ignoring the potential expense, the user’s data has to be sent off somewhere into the cloud, processed, and then the output downloaded. At any time, data can be syphoned off for use elsewhere, such as further AI model training. It’s not a secure way to work, especially in a corporate environment. Also, what if you’ve no Internet connection? The use of local AI technology means that you can create amazing images, audio, video, and text at any time and without exposing your data to a third party.
We are now at the foot of an exponential curve of increasing AI advancement that will likely change humanity forever, in much the same way as the Internet, and social media in particular, has. Intel, its partners and rivals are all investing in AI technology that will change the way we work, play and get around, likely sooner rather than later.