AI is on all kinds of devices now, including your TV. Companies like Samsung and LG are betting heavily on AI being the future of TV interactions.
AI TV features can mean very different things depending on whether you’re talking about image enhancements, upscaling, search, or even LLMs bolted onto the interface.
Some of this is mature and useful tech, but some of it is a solution looking for a problem.
Here’s how to navigate the AI feature list on most TVs today.
AI TV features that improve your content
AI image enhancements are helpful technology that can take your favourite content to the next level, improving image quality, detail, colours, and perceived motion.
AI Upscaling
AI upscaling has been a headline feature for years, and gives you a tangible increase in quality compared to older technology.
Older upscaling methods duplicate or average nearby pixels to fill the screen, making images look soft and blurry.
Modern AI scene detection identifies what’s on screen and makes upscaling decisions based on context. It uses machine learning to analyse the scene and reconstruct missing details.
The vast majority of streaming and broadcast television content is released in resolutions lower than 4K, meaning upscaling is important across the majority of media you watch.
AI Picture Processing
AI-powered picture enhancements are trained on image datasets and use these as a reference point for enhancing the content you’re currently watching.
HDR remastering adjusts the brightness and contrast of your content and boosts it to reach HDR output. It does so dynamically, adjusting output based on the actual image on screen rather than an arbitrary adjustment.
Likewise, colour booster enhancements adjust the white balance, saturation, and contrast of your content to produce rich but balanced tones.
AI Motion Enhancer aims to reduce the “soap opera effect” you may have noticed on older smooth motion features. If done well, motion enhancements reduce motion blur and flicker in fast-moving content, but if done poorly it can be jarring.
AI TV features you don’t really need
Alongside these helpful enhancements, companies have been prone to cramming AI in wherever they can. Here are some of the less useful features.
TV LLM Platforms
Platforms like Samsung Vision AI Companion that give you access to an LLM mean you can use services like Copilot and Perplexity on your TV, but there are few reasons to actually do so.
One use case companies push is that you can use the on-TV AI to perform searches without using a secondary device. For example, you could ask the TV to identify actors on screen while staying engaged in the content.
I reject the idea that performing a search while you’re watching a show means you’re still engaged in the content, whether that search happens on the TV or another device.
Another issue with the premise is that performing such a search on the screen is a great way to take everyone out of the moment if you’re watching in a group. The search will pop up on screen while you could have searched discreetly on your own device instead.
LLMs are primarily designed for one-on-one interaction, and putting one on a shared screen in the living room is an awkward fit.
There may be more uses that emerge as the technology matures, but right now this is a feature in search of a use case.
Generative Wallpaper
Samsung has placed Art Mode on more TVs across its range to provide a curated collection and liven up your living room with beautiful artworks from a range of artists.
The contrast between Art Store’s partnerships with artists like Sarrita King and generative wallpaper with no artistic provenance makes you wonder why the latter even exists.
Art Store features paintings and photographs from celebrated and emerging artists across a wide range of styles and aesthetics, making the image generation look a bit redundant even within Samsung’s own ecosystem.
AI Content Recommendations
Home screen recommendations surfaced by the onboard AI sound useful on paper, but streaming platforms already do this better with more complete data of your viewing habits.
The argument is that cross-platform recommendations serve up content you might otherwise miss, but it’s hard to know whether recommendations are actually personalised or weighted towards recommending content that the manufacturer has a commercial interest in promoting.
Content discovery is a problem that’s already been solved through the platforms you subscribe to.
Our take on AI TV features
The AI TV features worth paying for are the ones working in the background to elevate your viewing experience. Features like upscaling, picture processing, and automatic optimisations like Sports Mode and Game Mode produce noticeable improvements to image quality.
LLM features on your TV remain unproven and companies are yet to serve up a realistic use case for them. If you’re considering a new TV this year, these should be at the bottom of your list of reasons.
Comments