LG's custom chip is made to power AI in appliances and robots

The Neural Engine handles voice and image recognition, as well as security.

For the last few years LG has been pushing integration of AI features in all kinds of products with its ThinQ brand, and now it's designed custom hardware to enable on-device AI processing in "future robot vacuum cleaners, washing machines, refrigerator and even air conditioners."

Google and Apple both use custom AI hardware (Pixel Visual Core and the A12 Bionic's Neural Engine) in their phones, Tesla has unveiled its own processor to handle self-driving features and Amazon is reportedly working on a chip to help Alexa processing work offline. These "neural processing units" have had a lot of hype, but being able to handle complex tasks without a cloud connection could come in handy.


Similarly, the LG Neural Engine is intended to process image and voice data even when it's not connected to the cloud. That means a device with the chip could recognize its environment and adjust settings accordingly, navigate through a place it hasn't been before and understand a user's commands in natural language -- all offline. LG also claims powerful security is built in -- with all the data it takes in, resisting hackers who could remotely control a unit or use your washing machine's sensors to violate your privacy would an unwanted outcome.

The Korean company has been investing heavily in AI, most recently opening a new research lab in Toronto to follow its other North American location in Silicon Valley, so while this may be a first, it's definitely not the last thing we'll hear about LG and AI.