This article was taken from the June 2015 issue of WIRED magazine. Be the first to read WIRED's articles in print before they're posted online, and get your hands on loads of additional content by subscribing online.
While you're weathering a downpour on a Paris business trip, Stanley, your ficus at home in Surrey, sends you a text complaining that it is in dire need of a drink. You press a button, water is dispensed and Stanley is once more content.
We can now connect anything to the internet -- from pot plants to coffee machines -- thanks to the proliferation of "smart" objects.
But far from making such basic determinations ourselves, as is currently the case, the implicit promise from manufacturers is that we will soon live in environments that boast ambient intelligence in the form of technology that has independent decision-making skills.
Your future house will supposedly make the call on when to wake you in the morning, and then plan a nutritious breakfast in a valiant attempt to cancel out the pizza binge from the previous night.
The present smart-home market is fragmented and immature and will remain so -- even though "islands of interoperability will emerge around specific vendors and products", until at least 2020, Gartner's Nick Jones predicts. Research from the firm suggests that the fully mature smart home, which could include up to 500 connected devices, will not be a reality until 2022. As such, we must look beyond this slew of internet of things devices to understand what future smart homes will be like. "For every smart-home application you can think of, the first question that will need to be answered is who is operating it?" says Ramesh Kesanupalli, founder of the FIDO (fast identity online) Alliance, a not-for-profit industry consortium attempting to change the nature of online authentication. "You cannot have passwords," says Kesanupalli. "Passwords are the problem." Not only are passwords insecure, hard to remember and fiddly to input, but Symantec research that scrutinised 50 popular smart-home devices available in March 2015 revealed that none of them enforced mutual authentication or strong passwords.
Biometric authentication with a strong protocol behind it is consistently presented as the best alternative to passwords for keeping intruders from physically or virtually entering your home. By 2020, new biometric methods will displace passwords and even fingerprints across 80 per cent of the market, Gartner forecasts.
Backed by the likes of Google, Qualcomm, Alibaba and the Bank of America, the FIDO security protocol is designed to provide interoperability between biometric authentication tools such as ultrasonic fingerprint scanners, iris scanners or voice-recognition systems. There are no passwords at the back end; the value is tied up in your physical interaction with the authenticator.
FIDO is gaining traction in the technology world faster than any alternative, with Microsoft building it into Windows 10 and Samsung and PayPal already deploying it in high-end mobile devices.
Security is just one benefit of biometric authentication. Asserting your presence will come first, says Kesanupalli, "and then personalisation". If the products in your home know your identity, they will be able to automatically adjust to suit your preferences.
We are already starting to see these fields converge in the form of wearable products, many of which contain heart rate sensors and other biometric capabilities that make it possible determine your identity. One such wearable device, Nymi, has been designed specifically for security purposes, using your heartbeat to open your house and car doors.
According to Deloitte, wearables such as activity trackers will have a stack of capabilities that grow in complexity, with the basic ability to collect data at the bottom and, eventually, the potential to take "cognitive action" at the top. In the context of the smart home, this would translate to wearables feeding data to AI platforms that would develop pattern recognition and instruct devices in our homes to respond accordingly.
It will be the products that can use this personal data to participate in ambient decision-making that will ultimately unlock the potential of smart homes, believes Mark Spates, head of Logitech's smart- home platform and president of the Internet of Things Consortium.
Sensors are already capable of capturing plenty of data, from steps to sleep quality, location and mood. But the question, says Spates, is how do you use that data to make decisions on human behalf and make their lives better, more efficient?
Technology patented this year by Google under the name of "Security Scoring in a Smart-Sensored Home", will learn how long it takes you to leave the house in the morning, and will wake you up accordingly. Spates also points to Netflix, Facebook, Google Now and Nest as some basic examples of how machine learning and AI are starting to infiltrate our lives. But this behaviour-learning software has yet to be taught to respond to data acquired through sensors. "Even the best recommendation engines right now are still very elementary in their execution as software," says Spates.
What he envisages is a system in which your fan, your thermostat, your front door and your TV can interact about your presence and preferences, and theoretically then make decisions for you. "You're going to see that happen on the device level," Spates says. But to make this a reality, he argues that there must be more data points to work with. "The way to get more data points is you need more connected nodes in the home," he says. To determine the minimum requirements for a basic setup, Spates employs a theory he calls "the three-node rule". In essence, users will supposedly need at least three devices that can communicate with one another to form the basis of a smart home.
This is already happening, albeit in a basic form. For example, Misfit's wearables, light bulbs and Beddit sensors, can all interact with one another. Factoring in Gartner's predictions for market security, one can see that for this to grow into a more sophisticated and properly integrated system, this may well take at least five years.
Many people already own their first smart-home item in the form of their smartphone. They might also have speakers with Spotify Connect built in that they don't consider smart-home products. Likewise, it's possible that they own connected products, such as smart TVs, but do not immediately connect them.
Consumers "won't have a choice" about whether or not the appliances they purchase are connected, says Spates. "At first it's going to be the high-end things. But then, before you know it, the $70 version is also going to be connected."
And once the consumer base has become accustomed to these devices, according to Electrolux CTO Jan Brockmann, not only will connected products influence consumer behaviour, consumers will influence the development of the products they already own, simply by using them.
From a manufacturer's point of view, Brockmann says, the upgradability of software on the fly, without the need for the user to plug a device into a computer to manually download a firmware update, will increase the life cycle of products as well as the longevity of the manufacturer's relationship with the consumer. This is a particular boon for the white-goods market, where hardware life cycles tend to be between eight to ten years.
Spates agrees; your robot vacuum cleaner could be told how to sense different surfaces and clean them more effectively. Or you could come down in the morning to discover your coffee machine has adjusted to suit the quality of water in your area.
Electrolux's ovens and fridges, for example, will soon come equipped with cameras. Initially for cooking, this will allow you to check, via your phone, how your food is browning -- but you will still need to tell your oven what it is cooking. As object recognition technology improves, however, Electrolux's vision is that your oven should be able to determine, unaided, what it is cooking and decide for itself when dinner is actually ready.
Similarly, a fridge will only be able to automate home food ordering if it is able to recognise exactly what it is already storing. "Once you have good visual object recognition, you can make the other reliable technology," says Brockmann. Currently we are at the stage whereby a smart fridge could use Google's Search By Image service to identify an object each time it was put into and taken out of a fridge, but the system takes up to five seconds to analyse an image and return a result. Too long. Accuracy is also currently a limiting factor, but this technology is improving rapidly.
However, if automation and ambient intelligence are just over the horizon, will consumers want to remain lords and ladies of their smart homes, or cede control entirely to the machines? "For the mass market, they don't want a say," says Spates. He believes that when we buy a new connected product we will not want to sit down and program instructions into a ruling engine, and cites the example of the number of people that never even get around to programming the clock on their Blu-ray player. Instead, just like Nest, a product should connect, integrate and get learning straight out of the box.
"The products that are starting to gain traction are the ones that learn on their own and make decisions." The only decisions people will want to make, he says, are questions that can be answered "yes" or "no".
Your smart home might notice you have changed your morning routine, as you've started getting up early to do yoga. It's response? To ask if you want to adjust your coffee machine, alarm and light settings to match that routine for the rest of the week. "This is control," says Spates. "But it's control that comes after you are presented with a solution."
He does concede, however, that there will be a place for apps such as IF, which acts as a pared-back programming language for your home. As with IF, these might allow you to develop your own rule recipes -- "if this, then that" -- or they will create bespoke experiences on your behalf and perhaps act as a bridge, allowing previously incompatible devices to connect to each other.
Ensuring interoperability across ecosystems will be vital to consumer adoption. Common standards such as those being introduced among products by coalitions such as the AllSeen Alliance, of which Samsung, LG, Qualcomm, Sharp, Panasonic, Sony and Cisco are all members, are one answer. But, as outlined in The smart-home space race (see p3), getting such behemoth brands to work together has not been easy.
Ultimately, to make this brave new world a connected one, we are going to have to trust that these companies can play nicely and share their toys.
This article was originally published by WIRED UK
