What Are AI PCs? A Brief Explanation

Before I start with the explanation below I'd like to say that I believe that AI (artificial intelligence) as we are defining it today is not what most people envision as true 'artificial intelligence.' I look at it more as an augmented intelligence, one that uses your own interactions and those of users like you to generate best case answers, images, videos ect. We've been using the same software, same search algorithms, same voice commands, same pretty much everything for several years now. We never called it AI before because of the marketing strategy. Today we are just now seeing the marketing hype behind those same tools that have been learning our habits for years (I won't get started on the privacy issues here).

AI PCs - Intel's New Marketing Term

To put this simply the term AI PC is an Intel derived marketing term that requires specific Intel hardware and standards to apply. This doesn't mean that other PCs or devices can't run AI or that they aren't AI ready or AI chip equipped. So don't be confused here! 

Intel has stated that for a PC to receive the coveted "AI" label, it needs four things: a neural processing unit (NPU), a graphics processing unit (GPU), and the ability to handle Vector Neural Network Instructions (VNNI) and DP4a instructions (so that the GPU can handle video processing)  In other words what Intel is saying is that in order for them to call a new computer an AI PC it must be one with their latest CPU.

 

What is AI Ready and AI equipped

When it comes to slapping the AI ready label on things these are the few small things to know. AI apps and tools are currently used in two different ways. The difference between the two and how you are using them is really the difference in hardware requirements and what is needed. Basically any PC, smartphone, tablet or device can access and use cloud based AI.  On device generative tools require much better, more efficient hardware and that is were some of the confusion starts for most people.

One of the easiest ways to explain the hardware is to say the modern CPU as most people know it has become extremely outdated. New software changes take advantage of and use the GPU and a new NPU (secondary chip) to do the heavy work. These are both considered AI accelerators, deep learning processor, or neural processing units (NPUs). They have been specifically designed to accelerate artificial intelligence and machine learning applications and algorithms. (More on the differences here)

Tools like ChatGPT, Google image generator Gemini, or Microsoft CoPilot are all cloud based and do all their work in a cloud based system. This allows EVERY device to utilize AI because all the work is done by other, much more powerful systems, that then send the results to your PC or phone. This will soon change however as Intel has confirmed CoPilot and other applications will soon run locally on PCs and require better hardware to do so. 

PCs meeting those requirements are already shipping and other devices like the Apple iPhone, Google Pixel 8 and Samsung Galaxy phones are already using secondary more AI specific NPU chips for on-device AI.

In conclusion - don't be confused or fooled!
I'm writing this post I'm really hoping to take some confusion out of buying or upgrading your systems and hope that people aren't fooled into think they have to run out and search only for an AI PC to run their newly hyped apps. Yes newer hardware will eventually be required, and yes it only makes sense to adopt that hardware if you are already upgrading. No, that doesn't mean you will be left out if you don't!