Speech is silver, silence is golden.

Did you know that Siri, Alexa and other voice assistants in their “columns” are constantly listening and recording you in the background, and even storing your phrases? Even worse, they store a huge amount of data about your habits and preferences. The voice, for example, can easily be imitated, and if a person’s necessary set of phrases is available, it is easy to take control of all the connected gadgets without your knowledge. Hackers do not sleep, and it seems we have already overslept a small revolution. Spoiler: Our thoughts cannot be copied, so, the solution exists — please read till the end.

Having a microphone always switched on can lead to the leakage of personal and corporate secrets into the hands of scammers. For example, you might accidentally say aloud the details of your bank card, or make purchases in an online store. Modern digital devices work quite well in difficult conditions, like noisy rooms, and are capable of hearing and digitizing important information, even if music is playing. But voice recognition doesn’t always work flawlessly. In 2016, the Amazon smart column was in the midst of a scandal, when at the request of a boy named William to find him a “Digger Digger” children’s song, the Alexa voice assistant heard the request incorrectly and read the results of a search for porno, responding with the most vulgar adult words.

Moreover, this technology is far from perfect, because voice assistants are not able to accurately distinguish and recognize the voices of different people and can easily react to someone else’s voice command. Of course, if you are very limited in using voice assistants for a simple search on the Internet, then do not worry much. However, if you want, for example, to use the voice commands to confirm financial transactions or control the ‘smart home’ security system, the results can be very deplorable. It is possible that a technology will reach the level of development in the future to reliably identify the voice of a particular person and to make authorized access to devices or services for only a narrow circle of people.
In this regard, it seems that the technology of mental readout -neurointerfaces — is the most secure and protected channel of interaction with electronic devices, since it eliminates voice interaction. The user generates the commands mentally so they are unique and cannot be falsified. In this sense, each device interacting with the neurointerface, is “calibrated’” strictly for its owner.

By the way, let’s consider the risks of the new idea from Yandex. The company is developing an initiative to introduce a proper voice platform in the Drive carsharing cars, where passengers will be given an opportunity to use voice commands to launch some Yandex services such as weather, music, navigation and to answer incoming calls. This approach to the placement of the voice platform also carries hidden threats related to imperfections in voice technology. Remember when the system read out the incorrect search results for porno to a child? The system, with incorrect recognition, can easily create the wrong route and disorient the driver. The sudden activation of loud music or an incoming call can cause an accident. Neural interfaces, all other things being equal, exclude the errors of technical voice support systems, allowing to fully use non-verbal communication and control different channels. They can control a multimedia system, as well as securely manage incoming calls, which provides a high level of control security, for both the driver and passengers, and in general for everyone on the road.
BasisNeuro’s neurointerface is fully supported by Android-based, on-board computers that are integrated into modern cars, and can serve as a good substitute for voice assistants.


image: http://audioveda.info/wp-content/uploads/2013/10/2575194852_fc58c96c21_z.jpg