AI – Perspective and Overview of Applications

Artificial intelligence in consumer and home electronics

In the competition for the most popular buzzword of 2019, the two-letter combination AI will mostly likely lead by a long way. AI stands for artificial intelligence, which not only has what it takes to win that competition, but is also controversial enough to polarize the beneficiaries, in this case, the consumers.

gfu recently surveyed 2,000 households in Germany in 2019, and the results of this study showed that a large majority of over 71% are fundamentally skeptical about artificial intelligence. Although some can see the benefits – 58% of those surveyed believe that AI will make their work lives easier, whereas only 17% don’t believe it will help, and 25% are as yet undecided. Their fears outweigh their concerns when it comes to AI in the workplace, where 64% expect that the increasing use of AI will lead to job losses.

When you ask these sceptics why they tend to have such a negative opinion of AI, they usually list three reasons:

  • 60% explain that they do not want to disclose their personal information or share their behavior patterns.
  • 59% don’t consider that the technology is mature enough yet.
  • 56% fear that the AI technology will monitor and control them.

In contrast to the results of this study, IFA 2019 demonstrated just how extensively AI has already found its way into the products produced by the consumer and home electronics industry. That the topic is perceived as more negative than positive in Germany is mostly likely due, amongst other things, to the fact that consumers are not even aware of some artificial intelligence applications – AI technology often runs undetected in the background.

AI has been around for a long time

In most cases, consumers are not even aware that highly intelligent technology is at work in the background when they use their consumer and home electronics products every day. For example, if you take a photo with any current high-end smartphone, the excellent photo quality and the autodetection for optimal settings based on subject and lighting conditions is based on many thousands of reference photos. Without wanting to offend our readers, the outstanding photography and selection of the optimal settings is increasingly thanks to the artificial rather than the human intelligence.

Television sets use a similar technology. Screen sizes are constantly getting larger, and the number of pixels is increasing in parallel. 4K TVs – with four times the number of pixels compared to a HDTV – and in recent months, 8K screens, are becoming more popular. The only thing missing is a television broadcast that supports this level of quality. Traditional TV programs are broadcast in HD, at best. 4K, and especially 8K, are still considered exotic. That’s why TV manufacturers use upscaling technology. The two million or so pixels that were broadcast are upscaled to the eight million required for 4K (or 32 million for 8K). Originally, static algorithms were used to upscale, but AI is being used regularly these days. The technology compares the current image with millions of reference patterns, then selects an appropriate calculation for each scene based on those patterns. The AI technology doesn’t rest: it is constantly learning and working towards the optimal calculation. The result is an extremely detailed picture, where only the most highly-trained eyes can tell whether it is as it was originally broadcast, or whether it was upscaled to 4K or 8K.

It’s not just the video that is being continuously analyzed. AI is being increasingly used in TVs and audio receivers to improve dialog and reproduce surround sound, for example. For watching TV, AI can also help viewers find the content they will enjoy out of the growing library of available programs, provided that the television set is permitted to learn their preferences.

Success through trust

AI is a frequent companion, and not just for language assistants in the car, or Alexa, Google, Siri, and co. Machine learning and AI are working unobtrusively every day in the devices that surround us. AI technology moved into household appliances long ago. Washing machines learn, analyze the load of laundry, then select the appropriate program. This simplifies usage and saves resources, like water, detergent and energy, at the same time. And some of the increasingly popular robotic vacuum cleaners strive to optimize their results, for example, by remembering where crumbs fall more regularly to the floor from a table.

If the majority of those surveyed by gfu in this study believe that the technology is not yet mature, they are making an important point. After all, AI technology only “matures” when it is used. As it learns, it gets better and better over time. It’s how a language assistant moves from an “unfortunately, I didn’t understand that” response to a mumbled command, to a commentary-free execution of that same command over time. The prerequisite is that the user allows the technology to learn. It is only when personal data and behavior patterns can be analyzed, that the AI can improve.

“It is the same with AI as it is with many new technologies: Initially, skepticism prevails. Only when the benefits are recognized to be greater and outweigh the concerns, do people see it in a more positive light,” says Hans-Joachim Kamp, chairman of the gfu supervisory board. “AI is making its way into more and more devices. The speed of dissemination depends on the trust that consumers place in the technology. It is up to us, the industry, not to break this trust.”