The fact that modern gadgets collect information about their owners has long been no surprise. It is also not surprising that “voice assistants” – programs built on the recognition of human speech – have joined the surveillance of users. And if the products of Western IT giants usually use the obtained data to advertise goods and services, then the super-successful development of the Chinese corporation iFlytek has learned to send the overheard directly to the government of the PRC. It is the American side, whose developers have mastered the techniques of espionage no worse than their Chinese counterparts, the loudest about this fact.
I really hope they ban TikTok. The only people I want spying on everything I do are Google, Facebook, Twitter, Instagram, the Government, Apple, Amazon, every website, my FitBit, every app…
— Jason Mustian (@jasonmustian) September 18, 2020
“Do you remember how it all started”
A study of China’s path to mastering one of the most promising IT technologies was published by Wired, which named the founder of Chinese speech recognition Julian Chen – an employee of the American digital giant IBM , polyglot, Doctor of Columbia University and a martyr of the Chinese Cultural Revolution.
The fact is that one of the first programs that allowed dictating messages without touching the keyboard was developed by IBM, and in 1994 the company started looking for someone who would “teach” its product the Chinese language. The ambitious task was entrusted to Chen, who immediately realized that this innovation would completely change communication in his home country: in the computer era, Chinese writing was a unique problem for a long time, since there was no obvious way to enter more than 50,000 characters on a QWERTY keyboard.
To create his dictation mechanism, Chen broke down the most common Peking dialect of Chinese into tiny elements called phonemes. He then hired 54 native speakers and recorded how they read articles from Renmin Jibao, the main publication of the Celestial Empire. The IBM Research Laboratory in Beijing also contributed. Two years later, the IT specialist triumphantly presented his development in China. Even then-leader of the Chinese Communist Party, Jiang Zemin, participated in the demonstration of the program, and soon personal computer manufacturers throughout China, including competitors IBM, began to preinstall this prototype of voice assistants on their devices.
Distinguished Followers and Their Paths
Among those who were inspired by the triumph of IBM and Julian Chen was Liu Qingfeng, a 26-year-old PhD student in the Speech Recognition Laboratory at the prestigious Hefei University of Science and Technology. And in 1999 he founded the iFlytek company. Despite the skepticism of American competitors, Qingfeng was not afraid to challenge them under the slogan “The voice is the basis of culture and the symbol of the nation.”
The future Chinese IT giant launched its main development, the iFlytek Input application, in 2010. At the same time, Apple acquired Siri, a talking digital concierge who could answer questions from the iPhone owner. The functionality of the Chinese assistant was immediately focused on decoding the human voice.
It turned out that Siri and Input were the first swallows of voice interface technology. They were later joined by Microsoft’s Cortana, Amazon’s Alexa, and Google Assistant. At the same time, the Chinese application immediately had an important distinctive feature – a data confidentiality agreement, which allows it to collect and use personal information of users for “national security” purposes without their consent. This nuance in the activities of iFlytek, which owns 70 percent of the Chinese voice communications market with 700 million end-users, has subsequently been keenly interested in the Western press.
Consumer services are important to iFlytek, Wired reported, but about 60 percent of the company’s profits come from what the company’s 2019 report describes as “government-subsidized projects”. These projects include an “Intelligent Crime Investigation Assistance System” and support for big data technologies for the Shanghai government. In support of this fact, scientists from the Institute for the Future of Humanity at Oxford University, who study AI management in China, immediately discovered Liu Qingfeng’s connections with the Chinese authorities.
An impressive part of the study published by the American edition is devoted to the relationship of the Chinese IT giant with the Chinese government. But is everything so smooth for global digital companies from the United States itself?
At the epicenter of scandals
No matter how hard American corporations try to cast a shadow over competitors from China who threaten their hegemony in the global high-tech market, the voice assistants they have developed find themselves in the center of spy scandals more often. So, recently it turned out that Apple practices the covert inclusion of the already mentioned Siri in order to record user conversations, even when the program is not activated for them. And this violation of privacy is practiced by the Apple company systematically.
Earlier, it was reported on the Global Web that Windows 10 came with pre-installed settings that gave Microsoft the ability to track user-entered text and voice information, not to mention location, contacts, calendar entries and browser history. This data was probably used by the IT giant to display targeted ads, but the source also did not rule out that Microsoft is “looking” into user files at the request of the US ruling regime, although the developer’s “privacy policy” does not explicitly imply this.
In turn, the contractor Google VRT NWS in July 2019 “leaked” one of the Belgian publications a thousand records of voice assistants. According to them, the journalists were able to identify several people. At the same time, the Apple contractor admitted to the British The Guardian that on the records at his disposal, you can often hear how people have sex, buy drugs or describe their health. This also happens when people accidentally press the call button of the voice assistant.
In August 2019, Apple and Google announced that their contractors would stop listening to users ‘conversations with voice assistants Siri and Google Assistant, and that their employees listen to recordings of users’ conversations with devices only to improve the quality of their work. Whether this is true or not, remains unknown.
Miracle or bad
Voice assistants, like other digital services, have long become a routine part of everyday life. But in order to interact with us as efficiently as possible, applications need to recognize us. And in principle, there is nothing reprehensible and dangerous in the collection of information about users for training algorithms, if it is carried out in full accordance with the law, with the consent of the user, and also if the data is provided with adequate security. As the experience of Chinese and American services shows, ensuring privacy is not among their priorities. But even knowing this, people will not stop using comfortable and familiar gadgets. Stop using mobile apps today is like voluntarily giving up cutlery and eating with your hands: inconvenient, out of date, out of fashion.
It just so happened that tech giants were the first to realize that data is the new oil. And in order to maintain and increase their leading position in the global digital market, they will strive to obtain information about users at any cost. And illegal espionage (for any, even supposedly humane) purposes, no matter how cynical, is just one of their tools. No more.
The main question is – if companies are ready to do anything for the sake of data, then who will protect the user from total arbitrariness? Who should you go to for help? Paradoxically, only to the state. As the world experience shows, it is the state structures that are the only ones who can compel the tech giants to comply with the rules of the game and human rights. But the authorities must have a legal basis for this.
In Russia, the right of a citizen to the security of personal information from this summer is guaranteed by the basic law. Perhaps this fact will be unusual for us for a long time, and for someone it will be unpleasant. Exactly until it turns out that your Siri diligently records all your adventures. For unclear purposes.
