The all-singing, all-dancing, all-connected smart home has been getting gradually closer to becoming a pratical, everyday reality since the first smart devices began appearing a few years ago.
All the components are there – the hardware, the software, and the use cases. And while issues remain, such as interoperability, cost, and security, many of these will be remedied by the ongoing rollout of Matter, the Connected Standards Alliance’s (CSA) new communication standard.
However, there’s another pretty major issue: smart homes just aren’t very smart right now. Even with significant improvements to how our smart home devices communicate, there’s a lot of manual labor involved in setting up and maintaining a smart home.
Andy Watson, Director of Product Management and IoT Lead at Rightpoint doesn’t think that’ll be the case for long, though, based on his experience in implementing IoT solutions in a wide array of settings including smart homes. In fact, he gives it two years until the smart home finally gets its moment in the sun – and it’s all down to AI.
Slow and steady no more
Many of us have, without really thinking about it, already begun creating smart homes, simply by investing in one of the best smart TVs or best smart speakers. Some of us have gone further, perhaps by investing in one of the best robot vacuums or the best smart lights.
But by and large, it’s a far smaller group of smart home evangelists who have actually put time and thought into the building and operating their smart home setup - fewer still have done so extensively.
While devices like Amazon Alexa and Google Assistant have introduced AI-powered voice control and automation at a very basic level into our homes, Watson says these are barely scratching the surface of what smart homes will be capable of.
"In the 1980s, Steve Jobs said ‘for me, a computer has always been a bicycle of the mind - something that takes us far beyond our inherent abilities.’ Generative AI, in my opinion, just turns that bicycle into a motorcycle," he says.
People have traditionally used voice assistants in their homes to get information like the latest weather, or to turn lights on and off – it’s all very transactional. Generative AI has the potential to turn that relationship from transactional to collaborative, by making inferences by learning how we behave in our home, both on and off devices,, and feeding that knowledge back into smart home control systems.
In the 1980s, Steve Jobs said ‘for me, a computer has always been a bicycle of the mind - something that takes us far beyond our inherent abilities.’ Generative AI, in my opinion, just turns that bicycle into a motorcycle.
Andy Watson
The key, Watson explains, is in the immense sets of data generated by smart home devices: ”It can give you answers to questions in detail you didn't even know possible. It can help identify patterns and complex data and take action on those inferences. You generate a ton of data around the home, but it’s rarely used in concert.”
Of course, this isn’t technically new; Alexa, for example, develops 'hunches' for new automations and efficiencies based on your activity around the home. However, these are only skin-deep in reality – hunches can be disruptive, according to customer reviews, and they don’t seem very capable of personalizing homes with more than one resident.
Watson sees generative AI as a potential solution to improving smart home personalization. “You can really help situate generative AI into a specific way of thinking and act based on that, all this through normal conversations,” he says. In particular, he adds, AI will be a boon for non-technical users, making smart homes a lot more accessible.
Matter’s moment to shine
Evangelizing for both the smart home and AI is nothing new – I’d certainly hold my hands up to waxing lyrical about these technologies – but Watson truly believes we’re only a few years away from generative AI being applied to smart homes. “Generative AI is exponential," he says. "We're getting updates every single day about new capabilities or capabilities built on previous capabilities.”
So, how does Matter fit into this future?
While interoperability is becoming less of an issue as more Matter-compatible devices appear, AI could help to make controlling your smart home an even more seamless user experience – and Watson says this is only the beginning of how the technology can enhance a Matter-enabled smart home.
"Matter-based smart homes make more and more data accessible, which generative AI can really unlock the potential of, creating a much more tailored smart experience,” he explains.
“Even if you've got products that are supposed to work with one another, users still have to have the knowledge, the patience, the foresight, to manually configure their devices to work with one another to create those “automations”. These aren't really automated, but rather If This Then That configurations that you yourself set that you need to manually adjust.”
Better now than never
If we really are so close to the advent of the truly smart home, however, we might be in for a rude awakening if it arrives before we’ve had time to properly prepare; see, for example, the aftershocks that have followed this year’s boom in generative AI software and solutions.
"We're closer than we've ever been [to true smart homes]. Smart home technology has been around technically speaking for, you know, over 25 years at this point, but really, it's been more of a connected home than a smart home… I think that generative AI is going to push the industry to truly be a smart home industry.
“It's crucial for us to be talking about this in the smart home right now because there's so much to consider around how this will play out,” Watson continues. In particular, he says, there are challenges we need to navigate with care surrounding personal data and its uses.
“There's a lot to unpack there around the privacy concerns," he says. "You don’t want your personal data to be in the public domain, even if it’s a little difficult to trace.”
It's been more of a connected home than a smart home… I think that generative AI is going to push the industry to truly be a smart home industry.
Andy Watson
As of right now, your conversations on platforms like ChatGPT and Character.AI aren’t really private; while they can be obfuscated from your personal data, they can still be used to train the AI in the future; and of course, there’s always a chance that actors will find a way to access your data.
Watson suggests that “personal data generated from these smart home interactions need to live and be processed locally, either on your device or on your network to ensure it remains private.”
All quiet on the Western front
Generative AI is already being used in a wide variety of use cases in all kinds of settings, but certain opportunities are shrouded in uncertainty, largely due to a lack of guidance, regulation, and security. For now, it’s unclear how the big players in voice and AI are going to navigate this nebulous yet exciting prospect for smart homes.
The key players – Amazon, Apple, Google, Samsung, and the CSA – have been fairly tight-lipped so far, which is understandable given the complexity of the task at hand and the newness of both Matter and generative AI.
Still, it’s undoubtedly the case that these conversations are happening at a high level. The question isn’t if AI will see our smart home dreams become reality – it’s a question of when and how.
