Consumer electronics and building services that listen to your word - until recently that sounded like science fiction, but it has long been a reality and is developing into a real boom market. And in fact the idea is really cool. Instead of fiddling with a bunch of remote controls or finding the right one among half a dozen apps, you can simply control all of your building services with normal voice instructions. Whether it is about dimming the lights, regulating the heating or simply playing a certain song over the system - as long as the voice recognition understands the instruction and there is a connection to the respective technology, none of this is a problem. It is also very easy to call up information from the Internet and have it read aloud, for example the weather report for tomorrow or the traffic situation on the way home.
The race is on
Amazon's Echo is currently the top dog among smart home products - and is obviously selling extremely well. Amazon has been marketing its "Alexa" voice control technology in the Echo smart home speaker and the smaller Echo Dot version since last year. And apparently so successfully that in Germany you first have to be put on a waiting list in order to receive an invitation (!) To order - at some point. The PC giant Lenovo (formerly known as IBM) has just announced its own products at CES that will also use Alexa. And there are very understandable reasons why Amazon is pushing so intensely into this market. No other provider can interlink the areas of search, service and sales as closely as the powerful online department store, which is preparing to become Google's biggest competitor. Lenovo also relies on Amazon's Alexa technology and presented its own smart products at CES. Google itself naturally wants to counter this and has recently started offering a very similar approach with the "Google Home". Apple seemed to be ahead of the game for a while with Home Kit, but the system isn't much more than a nice idea. The Korean electronics giant LG and the German-Swiss company digitalSTROM are currently thinking at least one step further. Both presented voice-controlled robots in Las Vegas that should have what it takes to support us as digital butlers in everyday life. Google simply calls its Alexa competition "Home"
Hitech at its finest
For many a strange idea: Thanks to modern speech recognition, smart robots like the "Pepper" from digitalSTROM should soon become everyday companions The technology behind the various offers is indeed fascinating. Understanding natural language is still a gigantic challenge for every computer and has only become possible in its current form thanks to recent developments in the field of artificial intelligence (AI). The main difference between an AI system and a conventional computer is that it is self-learning. So it no longer needs human programmers to implement new voice commands. Rather, the system becomes a little smarter with every request and can therefore, in contrast to the unintentionally funny beginnings of speech recognition, increasingly understand dialects, accents and colloquial formulations. Speech recognition has its pitfalls ... not just for Scots! This independent learning requires considerable computing power and, above all, requires gigantic databases that are growing at breakneck speed. Neither this storage capacity nor the necessary computing power are available on a cheap device for home use, which is why the voice inputs are sent over the Internet to large servers, where they are processed, interpreted and sent back to the device at home. And this is exactly where many see a danger.
Attentive listeners
Because so that they can react immediately to voice commands, these devices constantly eavesdrop on their surroundings - Amazon's Echo, for example, uses seven individual microphones. If you recognize a preset keyword ("Alexa", "Okay, Google", "Hey, Siri" ...) record the following voice instruction and send it to the server on the Internet for analysis. So far, so good and helpful, but nobody really knows what exactly happens to the other sound recordings in the area. Amazon states, for example, that the last 60 seconds of the recording are stored internally in the device, after which the data should be overwritten and thus permanently deleted. This relatively long recording time is necessary in order to correctly interpret the actual voice commands in the context. Other manufacturers are silent on such details, but will probably work with similar methods and data. With no fewer than 7 microphones, Amazon's echo constantly listens to his cue. If it stayed that way, there would actually be little to criticize about Echo & Co. It may seem a little strange to be overheard by an electronic device all the time. But if the data only remains in the device and is deleted again after 60 seconds, this invasion of privacy is hardly worth mentioning. Every Facebook user, every owner of a smartphone voluntarily reveals a lot more about himself and even explicitly allows the provider to use this personal data for commercial purposes of all kinds. But a little uneasiness remains. First of all, a certain mistrust of the technology companies gnaws at the good feeling. Speech recognition, everyone agrees, will be one of THE key technologies of the next few years and open up a gigantic market. And the better your own speech recognition works, the bigger the piece of cake that the provider can cut off from the expected sales pie. And since the quality of the recognition depends largely on the learning abilities of the AI server, it seems quite tempting to use the constantly listening devices clandestinely to build up your own language database and thus gain an advantage over the competition. Data protection officers, hackers and security experts all over the world will be lurking to prove such an improper use of the audio data to a manufacturer. But as you can currently see from the example of Volkswagen, moral and legal concerns can go overboard in tough competition if the pressure of competition is great enough. So at least it cannot be ruled out that large providers such as Apple, Amazon or Google will succumb to this temptation at some point. The terms and conditions, which one must accept before using such a device, allow the provider in any case, as usual, an enormously extensive use of the data collected. There is also of course, at least in theory, the risk that an outsider will gain access to this data or directly access the built-in microphones. In order to be able to function, devices with voice technology obviously need two-way access to the Internet - they must be able to both send and receive data. Even if the manufacturers credibly assure that they have taken all conceivable safety precautions in this regard, such protection can never be 100 percent secure. So it is not entirely out of the question that the domestic smart home loudspeaker could be turned into a bug by a hacker. And the secret services around the world are certainly looking for ways to make the new technologies usable for large-scale eavesdropping. Are these all reasons for exaggerated paranoia? After all, nowadays we are constantly surrounded by devices such as smartphones, tablets or laptops, which are much easier to hack and which we voluntarily leave behind an unmistakable data trail. But with the speech recognition and the necessary constant eavesdropping by the smart devices, the whole thing takes on a whole new dimension. Imagine talking to your partner about planning your dinner and suddenly being interrupted by the intelligent building services with a reference to a new pizza delivery service nearby. It's a creepy idea today, but ten years from now it may seem perfectly normal. Because Apple, Google, Amazon, Facebook, WhatsApp, Twitter, Instagram and many more have long since proven: Despite all the original concerns, people are ultimately willing to reveal their personal data if only the benefit is right for them. Smart building technology with real, intelligent speech recognition will come, no question about that. But maybe now is the right time to think about which freedoms you want to give the developers of such technologies and which you prefer not.
