What is TV AI Artificial Intelligence?

It's less about intelligence and more about learning

by Steve Withers Mar 8, 2019 at 7:01 AM


  • Manufacturers seem to be adding artificial intelligence to everything these days, and it's only a matter of time before your toaster is AI-enhanced. TVs aren't immune to this trend, but how does AI work in a television and what are the benefits?
    When we hear the words 'artificial intelligence' most of us probably think of Skynet becoming self-aware in the Terminator movies. That's certainly the classical definition of AI: a machine that learns and understands to the point where it can pass the Turing Test. To do that you would have to be able to hold a conversation where you were unable to decide if who you were taking to was a human or a machine.

    The AI that is employed in today's TVs is a far cry from HAL9000, the murderous computer in 2001: A Space Odyssey. In reality the kinds of technologies that the manufacturers are really referring to are machine learning algorithms, but let's be honest that doesn't sound as cool as AI.

    When it comes to AI in TVs, there are three main areas where such learning algorithms are being employed: image processing; sound quality; and smart features.

    Image Processing
    In terms of image processing, modern chipsets are capable of making a staggering number of calculations in real time. This is generally applied to upscaling, edge enhancement and noise reduction features, where the processing takes a lower resolution image, cleans it up and then fits it to the native resolution of the panel.

    Since this involves calculations based on millions of pixels it requires a serious amount of processing power, but manufacturers also use databases of images to improve their upscaling, edge enhancement and noise reduction features. The larger the database, the better the processing.

    This is where AI comes in because if the processing can not only analyse the images that it is receiving, but also learn from them it can make the results better. The application of learning algorithms is obviously limited by the processing power actually built into the TV, but there is another approach.

    In the past the manufacturers would build an image database and then load it into their new TVs before shipping them to stores. Now they can build a database and use machine learning to expand it with new images that evolve over time.

    This database can then be periodically uploaded to the TV, ensuring the processing remains up-to-date. By applying the machine learning remotely, the manufacturer isn't limited to the amount of processing on the TV itself, which is sensible.

    Sound
    Manufacturers are also using AI to improve the sound quality of their TVs. AI sound enhancements use spatial awareness technology to analyse an environment and adapt the audio to create the best experience based on the sound settings and the location of the TV. The learning algorithms can also analyse the sound for specific scenes (news, sports, game etc.) in real time and adjust the audio accordingly.

    Previously manufacturers used metadata to apply sound processing in their TVs, but the addition of AI technology provides the ability to enhance and optimise audio based on analysing sounds in real time. The intended result of using AI enhancements with a TV's sound is to deliver a better experience for users, with audio that is optimised for the TV, the settings, the room, and the content.

    Smart TV
    The third area of a TV's performance where AI has been making a huge difference relates to its smart platform. There are a number of features that can benefit from learning algorithms, but the main one relates to content discovery and recommendations. The TV can analyse your viewing habits, learning which shows you enjoy, discovering new content, and recommending additional programming. It's a very useful feature and helps consumers deal with the overwhelming amount of viewing content that is available these days.

    This application of AI also ties in to how you actually interact with your TV, and we are now seeing extensive use of voice control. This can either be applied through a manufacturer's own AI assistant software such as Samsung's Bixby, or using third party versions like Google Assistant, Amazon Alexa, or Apple's Siri. The important aspect of these AI assistants is that they learn from interacting with you, ultimately making the experience more natural and conversational.

    There are other applications as well, but ultimately when it comes to AI Smart TV platforms it's all about tailoring the experience for individual users by monitoring their interaction with the system and customising it for their specific needs.

    AI or machine learning is gradually going to revolutionise how we interact with technology over the next few years, and hopefully it will ultimately be beneficial. There are certainly genuine concerns about just how much a company knows about our personal habits, and that may well require a degree of legislation. However, as far as AI is concerned you needn't worry, there's no chance of you TV suddenly becoming self-aware and starting World War Three.

    To comment on what you've read here, click the Discussion tab and post a reply.

    Share This Page


  1. This site uses cookies to help personalise content, tailor your experience and to keep you logged in if you register.
    By continuing to use this site, you are consenting to our use of cookies.
    Dismiss Notice