A refrigerator that can think: What's next for the homes of the future

Smart home technology is becoming increasingly important to both developers – through their apartment plans – and the Australian consumer – as a way to save on energy bills.

But for LG, the next step in smart homes is all about artificial intelligence (or AI), with its idea first outlined at the technology fair CES 2018 in Las Vegas, and with more details coming at the IFA Global Innovations Show.

Does this mean after years of being yelled at, your television will finally start yelling back? Well, not really – although if you ask your smart speaker nicely, it will sing for you.

The proposed idea is that while refrigerators, televisions, dishwashers, washing machines and lights could respond to voice commands, they might also learn from your daily habits to the extent that they’ll already know what to do before you tell them.

It won’t involve needing to learn a whole new operating system, either. Senior director of LG Global Communications, Ken Hong, said that the key to success was complementing existing technology, and embracing tech that buyers were already familiar with.

“We’re not competing with Google Assistant, we’re not competing with Amazon Alexa – we’re not trying to get consumers to adopt a whole new interface”, he said.

And, rather than attempting to lure customers with new gadgets each year, the focus would be on household items they’d want to keep.

• Related: Do you need a smart mirror?
• Related: Turning an existing home into a smart home
• Related: The benefits of living in a smart home

“The more you use it, the smarter it gets”, Mr Hong said. “You won’t want to replace it because it’s your partner – it knows what you like, it will predict what you are doing, and thinking.”

Smart televisions, refrigerators and ovens could all “talk” to each other, via a cloud system and work to benefit the user. But, he added, the challenge was ensuring maximum benefits while protecting the user’s privacy.

LG’s AI expert Peter Pyungchui Kim said there was security at every level of the AI process – the hardware itself, the connectivity that existed in the home, and the software the appliances used.

And he reiterated that the goal was to create appliances that actually increased in value over time.

“If you purchased a washing machine – and at the time of purchase, it spends a certain amount of energy per cycle [and] a certain amount of time”, Dr Kim said. ”

But as you go on, the amount of energy and time will reduce. That is what we want to deliver.”

This means that all these appliances need is a Wi-Fi connection. Dr Kim explained that the amount of number crunching and data analysis needed for the machine learning means more processing power was need than was feasible to sell in a household appliance.

There are still areas to improve, Dr Kim said. For one, it might be better if the device could interact with people in a more natural way – people have higher expectations from a device they can talk to, and have “higher”, more concept-based goals like comfort and security.

LG senior marketing manager Brad Reed said that the first DeepThinQ to reach the Australian market would a robotic vacuum cleaner, which he anticipated being available in October or November.

He said the goal was for it to be a mainstream product, rather than something that required tech expertise.

“It is a product that doesn’t need a lot of complex setup, doesn’t need a lot of integration. I would see that as being our philosophy moving into other products too,” Mr Reed said.

This would be followed by Wi-Fi-enabled washing machines and refrigerators next year.

“We’ll probably start to see some kind of AI technology in those products, but it’s a different kind of product. You’d be getting better functionally, better usage, better support”, he said.

The CLOi home robot, on the other hand, it unlikely to be available any time soon.

Looking long term, AI-powered devices could be geared towards saving money via electricity conservation.

“It definitely makes sense, but it requires some interrelationship with the energy providers” Mr Reed said. “It’s a pretty simple step to take with something like a dishwasher – it can learn when the tariffs are low and switch on then, for example.”

Nicole Frost travelled to South Korea courtesy of LG