Stacey on IoT | Internet of Things news and analysis
Internet of Things
by 5 Comments
Consumers are still not sold on the smart home, with their biggest concerns being privacy and fragmentation. That’s according to Raza Shah, technology strategy and design leader at Jabil, a contract manufacturer that has manufactured many smart home products.
Shah and I spoke this week in the wake of the release of Jabil’s 2023 Smart Home Report, which looks at the future ways consumers might interact with the smart home. It sheds light on how far off manufacturers think Matter deployment is, which sensors are the most exciting, and how AI will affect product development. For me it’s a chance to look at what’s coming next in the smart home and what we’re still struggling with.
I purchased my first smart home device in 2012. It was a Wemo smart plug (ironic, given this week’s podcast discussion) for $50, which I then used to turn my Christmas tree lights on and off using the app on my smartphone or a schedule. My husband was baffled by the idea since we already had an outlet with a timer for our holiday lights. He was unimpressed by the fact that I could turn the lights on from a restaurant or the car or our bed using just my phone.
At the time, most of the world was like my husband. Connected devices that added remote control were viewed as expensive toys. And as more companies started launching these devices, it became clear that mainstream users weren’t going to adopt smart home products unless they were cheaper, they did more, and it became less complicated to buy them. Even then only some devices worked with Google’s Nest thermostat, for example, or with specific smart hubs.
More than a decade later, most devices are still expensive (or require a subscription) and still don’t work together, despite Matter launching last year. And there still isn’t a compelling value proposition to connect all the gear in the home. Adding to this mix of uncertainties is concern over security and privacy that leaves consumers wondering if it’s even safe to connect home devices.
As Shah said to me: “The first struggle for consumers is the amount of choice and the fragmentations. The second struggle is believing that they are constantly being observed.”
Matter was designed to solve the first struggle, but even Shah admits that so far Matter has had a slow start, and he believes that the more complicated experiences — such as enabling users to control devices using multiple controllers, like Alexa and HomeKit, through the multi admin function — will take two to three years to occur. Despite that, the Jabil survey notes that 73% of companies plan to integrate Matter into their devices in the next 12 months.
“It’s a slow start, but it is a start,” said Shah. “For once, all of the giant ecosystems have come together, so of course there is a bit of jostling and shuffling of elbows.”
As amusing as it might be to imagine Apple and Google, or Samsung and Amazon, crammed into the back seat of a minivan, fighting about who’s crossing over the invisible line delineating each company’s “side,” it’s hell on consumers who have been eagerly awaiting real device interoperability. Based on Shah’s comments, if fragmentation is the biggest concern, smart home adoption isn’t coming until Matter actually works.
As to the second-biggest concern, in which consumers are focused on their privacy and how their data gets used (or hacked), we’re probably even further than two or three years out from solving that. Shah said that while Matter does provide some security, he thinks that “one of the principal challenges [consumers have] in embracing smart devices into their homes is they are aware that everything they are talking to is being listened to.”
Shah said this is partially a consumer education problem. He thinks manufacturers need to be aware of the types of data they collect, and engage with consumers early on around what gets collected and how it is used. He’s also “interested” in how the White House is approaching the cybersecurity label for IoT devices, and whether or not privacy could be part of that effort. He said efforts by the Connectivity Standard Alliance, the organization behind the Matter standard, to create a data privacy working group are also welcome. “It should be both a top-down and bottoms-up effort,” he said.
Jabil, which works with startups as well as with giant companies that are already focused on securing and developing privacy policies for their products, is also working on creating a service that would get data from medical devices into the cloud securely, all while protecting consumer privacy. The service has the benefit of protecting consumer privacy and also letting medical device companies bring products to market faster.
The report also touches briefly on the impact of AI on smart home products, with 23% of respondents labeling it as disruptive. Shah said AI will drive newer applications based on personalized data from individual smart homes, and even data extrapolated from multiple homes in a neighborhood. I think that’s possible, but I believe we should solve the privacy issue first.
Want the latest IoT news and analysis? Get my newsletter in your inbox every Friday.
Filed Under: Analysis, Featured
Enter your email address to receive notifications of new posts by email.
There is no reasonable, private solution to voice commands. To process a voice command you need about $20,000 worth of hardware for 0.1 seconds. And how many times a day do you use a voice command? Probably no more than 20-30. So you need about 3 seconds a day on this $20K piece of hardware.
You can do Alexa level processing privately in the home, but you need to leave something like an RTX4090 running 24/7 and the 4090 is going respond in 1 second instead of 0.1 so you are going to notice the lag. Not many people want to pay for the electricity to leave that 4090 running 24/7. Plus they cost $2000 and you have to manage this technical mess. And you are only going to use it about a minute a day.
So you are basically forced to a cloud solution unless your name is Musk and you don’t care about the price. If you are going to use someone else’s cloud you have to pay for it somehow. Either you fork over some cash, or you let them data mine.
Free and private don’t come together in the same package.
That’s definitely true for most natural language processing, but there are some options if you are willing to choose systems, where you have to say very specific phrases in a very specific way.
(Again, as someone who is quadriparetic, I am completely dependent on voice controls and follow this technology closely.)
1) Just as one example, Alexa echo models which are connected via Zigbee to local devices do offer very limited voice processing that doesn’t require an internet connection. I use these in my own home for my “plan C” when both regular Alexa cloud processing and Siri are unavailable. So I will have one light switch in each room in a pathway from my room to the front door available for this local voice control.
2) one of the oldest options is voice mouse navigation rather than voice control. This allows you to hands-free control a mouse moving around a PC or tablet screen, so you can control home automation that way. You can already get some local voice navigation on iOS devices, which allow you to move a “voice, mouse“ and actuate device tiles in a home automation app. Again, much more limited and tedious than just a regular smart cloudconnected, voice, assistant, but for those of us who need this, it is an option.
3) Home assistant has some recently released local voice control features, but my understanding is that they all require you to press a button to activate, so they don’t work for someone like me. (I have the same issue with the voice command TV remotes where you have to hold down a microphone button while you are speaking.) anyway, they are calling 2023 “the year of voice“ and looking very closely at this area.
4) The EU SIDN Fund has sponsored an interesting, local voice project called “Candle” which allows you to put together something like the raspberry pi setup you mentioned, but at a cost of under €200. Again, it’s not a full smart speaker replacement, and it is some work to set up, but it is an interesting approach to local voice control.
https://www.candlesmarthome.com/what-s-new-in-candle-2-0
Anyway, we did have some local voice control before echo arrived and there is still some now. Definitely not as fun or full featured as the off-the-shelf commercial cloud-based products, but there are some options for some people that won’t require spending thousands of dollars.
We should also mention that Amazon is telling developers they will be able to have local voice control of matter devices “when the Internet is unavailable“ if they build that into their device connections:
https://developer.amazon.com/en-US/docs/alexa/smarthome/matter-support.html
But I don’t think that’s here yet.
You need an Alexa with their custom AI chip in it for local commands to work. Only their top models have that. Go into settings for the device in the Alexa app and see if you have an option to turn it off or on.
The other big issue is that everyone wants a subscription to access their (actually, my device, I bought it) device. And force me to use their app. Looking at you Liftmaster.
The link to the MyQ app privacy policy and T&Cs fails – so I’m to sign up blind? When you can find them (and only on a phone, I could find a web accessible version) they are pages and pages long.
And incompatible with everything else I own.
Your email address will not be published. Required fields are marked *
This site uses Akismet to reduce spam. Learn how your comment data is processed.
What’s coming next in the smart home? – Stacey on IoT
RELATED ARTICLES
