Currently the CEO of AI chip company XMOS, Mark Lippett is an experienced business leader with over 25 years’ experience in technology.
getty
If you look at the biggest technological breakthroughs of the past 100 years, what do you think of straight away? Perhaps the invention of the electronic television? Or the invention of microwaves? Or of course, the rise of computers, the internet and the smartphone?
Whatever springs to mind at the moment, in about 20 years’ time, you might want to think about adding the “artificial intelligence of things” to your list.
What Is The AIoT?
The artificial intelligence of things (AIoT) promises to be one of the most exciting technology developments we’ve ever experienced. So much so, in fact, that it’s likely to become a $3 trillion industry by 2024.
Functioning as the convergence of artificial intelligence (AI) and the internet of things (IoT), the AIoT represents a new way of delivering AI — taking it out of the data centre and embedding it directly in the devices we are surrounded by daily.
This way of delivering AI is a game-changer for almost every industry, including the smart home, healthcare, transport, manufacturing and many more. The technological possibilities are almost endless — from traffic monitoring and security within smart cities to remote patient monitoring in healthcare and true personalization in smart homes. The most exciting part is that we probably haven’t even thought of the most powerful implementations of the AIoT in our society.
How Is The AIoT Such A Powerful Tool?
The reason the AIoT has so much potential is because of how it delivers AI. It fundamentally breaks away from the traditional cloud computing structure. By embedding intelligence directly within end devices themselves, whether it’s smart speakers in the home or smart traffic cameras in a city, we change everything.
First, devices don’t share your data with anonymous tech giants, compromising your current and future privacy. Second, devices can make almost instantaneous decisions because they don’t have to suffer the latency issues associated with sending data back and forth from the cloud. Thirdly, it means that smart devices can exist anywhere — they don’t suffer from absent or varying network connectivity problems, which is only likely to get worse in the coming years as the world connects more and more devices to the cloud.
Think about autonomous cars for example — an autonomous vehicle needs to be able to make instantaneous driving decisions based on the immediate surrounding environment. A traditional cloud-based AI model cannot support autonomous cars because of network connectivity issues, latency and bandwidth. The AIoT promises resolution of this issue and many similar challenges in an incredibly diverse range of electronics products.
The Catch
The success of the AIoT is completely dependent on the continued development of the technology underpinning it. While we can theorize about the impact it will have on society, there are still many drawbacks to the current generation of processors that prevent it from being introduced on a large scale.
The AIoT is a relatively new ecosystem, and it isn’t an easy feat to embed the computing power needed to run it in the small, low-power devices where it is required. For AI to be effective, it needs low-cost processors capable of high performance over an extended period of time to be able to work in almost any “smart” application. But existing chips tend to be big, pricey and designed for specific purposes — not making them a great choice for use within the mass manufacturing of an incredible variety of devices.
Nevertheless, the semiconductor industry is working hard to solve these issues, and over the past year, thanks to significant breakthroughs in computer architecture and semiconductor process technologies, the market has started to see the first truly flexible, low-cost, high-performance processors enter the fray. The feat of engineering that has gone into these processors cannot be understated, and they represent the foundation of the future AIoT.
What Businesses Need To Do To Take Advantage Of The AIoT
If you’re a business that sells electronics for the smart home, healthcare, transport, manufacturing, you can finally start planning to put AI at the heart of your products — especially if you are an electronics manufacturer for the smart home. The pandemic has placed an even bigger focus on the quality of life we have at home, and as such, the smart home represents the AIoT’s biggest immediate opportunity.
So what can you do to prepare yourself? Everything starts with the processor. Every electronics engineer will tell you that to embed AI into devices, you need to massively increase that device’s processing power — but without adding to the overall cost. It’s an age-old engineering challenge, but finally a surmountable one. The best advice is to look for future-proof chip technology — technology that’s high-performance, low-cost and flexible enough to be used in almost any AI application.
Using the right chip is one thing, but deciding what the end intelligence of your devices should look like is another. We’ve all seen the “technologification” of mundane things like toothbrushes, which end up being gimmicky rather than useful. Instead, think about how intelligence can replace or augment an existing process to make it more convenient. For example, if you’re a manufacturer of refrigerators, adding intelligence to your products to connect them to your customers’ Twitter account is hardly useful. But adding intelligence that analyses the contents of the refrigerator and automatically adds items to a shopping list when they’re running low represents valuable utility.
Once enough manufacturers address both of those challenges, the AIoT will become a real force. Time will tell, but perhaps adding AIoT to your “biggest tech breakthroughs in modern history” list will happen sooner than you think.
Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?
