Generative AI tools go way beyond making up jokes and generating text at your leisure. In fact, their reach and power could change how you control devices around your home, and what they can do for you.
Although it may seem like generative AI arrived recently, Andy Watson, director of product management and IoT lead at Rightpoint, says that isn't necessarily the case: "There have been folks working on this for decades, but, in terms of the everyday experience and interactions with it, it's a drastically new and powerful technology."
Yet for many of us, AI's greatest impact will reach us right where we live: in our homes. Through automations, virtual assistants, and machine-learning algorithms, AI can make homes more efficient and give you back valuable time.
Of course, AI technology is no newcomer to today's smart home. Whenever you talk to your smart speaker or activate your home security system, you use a resource developed with deep learning in its stack. And while we're a long way from the fully automated, futuristic vision of homes portrayed in "The Jetsons" cartoon show, our smart home journey can start with baby steps -- adding an Echo Speaker, Apple HomePod, Google Nest, or even a Samsung TV, installing a smart lightbulb or plug that you can control remotely, or using a security camera you can view on your phone.
However, the reach of AI in the smart home has the potential to go much deeper. Increased automation could enhance our personal lives and boost our productivity. Let's take a look.
AI runs deep in smart home tech
"Bringing the smarts into your entire home is a huge, huge benefit that generative AI can provide," Watson explains.
These benefits can extend beyond today's smart devices, for example, into intelligently enhancing security footage in real time, especially in low-light situations, or generating routines and automations, according to learned patterns and then giving a homeowner the option to confirm them. There's also the opportunity for generative AI to provide predictive maintenance.
AI can learn from your home, your appliances, and your consumption, making inferences from data that, Watson says, "help you identify items in your home that need specific preventative maintenance, either done by the homeowner or to be scheduled through the generative AI on your behalf with a trusted third party."
Just look at what Google Assistant can now do for Pixel smartphone owners. The virtual voice assistant can screen calls for users, speaking to the caller and notifying the user of what they're calling about, or simply declining to put them on the phone if it's a spam call. Generative AI tools in the smart home could evolve further to set up appointments for your HVAC maintenance, landscaping, or gutter cleaning, which would be based on your specific schedule and availability, using data gathered from the devices.
"If you have a smart hot-water heater that can connect to the internet and gather some data around it, the AI generator would be able to make inferences off of your usage," says Watson, and determine when maintenance would need to be performed.
Living in our fixer-upper house -- which has required a lot of intervention from different tradespeople -- I can imagine a future where I ditch my iPhone Reminders app in favor of a generative AI assistant.
This assistant could make an appointment with the maintenance company directly, provide me with trusted resources that my neighbors have highly rated, or it could give me personalized recommendations, based on the likes and dislikes it has learned about me. In short, generative AI could soon reduce the many decisions you have to make to just a couple.
What about ChatGPT and voice assistants?
Generative AI might help make the home of the future smarter, but the impact of AI can already be seen and heard. Amazon, Google, and Apple have been using natural language processing (NLP) systems to create the voice assistants we have at home, including Alexa, Google Assistant, and Siri -- and like Google Assistant, Alexa will also get a generative AI upgrade in the coming months.
Watson believes generative AI will turbo-charge these voice assistants: "Voice assistants today... People use them for very transactional things, 'I need something, and you give me something; turn on the light, get the weather,'" he explains. "But from the generative AI perspective, something like ChatGPT has the potential to turn that relationship from transactional to collaborative."
Instead of simple automations that need to be manually configured by the user, generative AI could instead suggest new, more intricate ways that a smart home can work based on what it knows about the user's behavior. This could include turning on a fall-detection system when a specific family member enters the room, or running customized automations when a family member is identified on a security camera.
This evolution is already evident in the ways that a smart HVAC system can use generative AI to optimize heating, air conditioning, and ventilation. The system combines data from user patterns and environmental factors, such as outside temperature and humidity.
"The real power that we see behind generative AI models lies in their ability to make inferences of immense datasets," Watson says, further explaining that all smart home products generate a vast amount of data rarely used in concert with one another.
Voice assistants, such as Alexa and Siri, gather user data to learn more about preferences and patterns. Your iPhone uses machine learning to recognize which speaker you typically play music on while cooking dinner, and starts making suggestions when it comes to that time of day, for example.
But this preference data could be used for much more. Watson explains that if generative AI models can infer answers from the user-behavior data they collect, they could make our homes smarter without needing manual intervention and foresight.
One company, Josh.AI, is leveraging this potential to add the power of generative AI to the smart home. The company officially launched JoshGPT, a voice assistant powered by OpenAI's GPT technology, that can go further than Alexa and Siri right now, thanks to generative AI. Beyond asking Alexa to give a weather report, Josh.AI users can ask JoshGPT to check the weather, turn on a light, and play music in one single request.
JoshGPT can generate answers to your queries, rather than just search for them online or repeat what websites say. You can ask JoshGPT questions much like ChatGPT, but through a voice conversation using microphones and speakers around your home.
And like JoshGPT, our smart home voice assistants could become smarter than ever and able to handle multiple tasks at a time, such as three questions in a row. "Generative AI can do the thinking for you, if you will, so that anyone can have a smart home," says Watson.
This is where Matter comes in, sort of
The potential of generative AI in the smart home comes at a time of significant advancement in connectivity standards.
One of the most important step changes in connectivity is Matter, an open-source standard developed with the influence and investment of big players, such as Google, Apple, and more. Matter brings seamless interoperability between smart home devices that weren't previously capable of communicating with each other.
Since Matter was launched in 2022, many smart home manufacturers have begun adopting the standard to make interoperability a given in their products.
Watson agrees that Matter makes it possible to have "interoperability between a slew of different devices across different verticals, and multi-admin support for essentially-agnostic voice assistants." Multi-admin is at the center of Matter's purpose, as it means different users can control all their Matter-enabled devices without choosing a single platform.
"With the consolidation of user experience and the interoperability between these different devices, we now have the potential for a generative AI model to have access to an immense dataset it never had before, to help personalize the smart home experience to the individual homeowner," Watson adds.
What about ethics?
Our homes are our havens of privacy, so the idea of an AI model using data gathered from devices placed throughout a residence won't appeal to everyone. For all the convenience and benefits of having more AI at home, the potential risks mean some could balk.
"AI is a very broad technology. It's fast-moving. I think that's why many people are concerned about the trajectory of where it could take us," Watson says.
He believes AI requires thoughtfulness, morality, ethics, and some regulations to ensure safety and effectiveness. "In almost everything that we do when it comes to social media, any product that you're not necessarily paying for, you are the product in that equation," he adds.
Watson says consumers will have a choice about whether to trade convenience for privacy. Will they, for example, be willing to share data about their behaviors with other companies to aid the creation of lower-cost products and services?
"Transparency is step one," says Watson. "People need to know how their data is being used, whether it is being stored, where it's being [sent]."
Different consumers will have different opinions, as with any technology. For now, I'll continue to dream of a world where generative AI can schedule home-maintenance appointments without me having to do anything, set smart home devices that are tailored to my behavior, and where Alexa and Siri can finally understand what I'm asking them -- at least 80% of the time.