Behind the booming sales of smart speakers, who will protect the privacy of users?
In recent years, domestic smart speakers have developed rapidly, and now they have entered a stage of rapid growth.
The statistics of the smart speaker market in the first half of 2019 show that the sales volume of the smart speaker market in China was 15.56 million units, a year-on-year increase of 233%; the sales volume was 3.01 billion yuan, a year-on-year increase of 149%.
Smart speakers have been recognized by the public for convenience and interactivity, and potential information security issues have gradually been exposed.
In April this year, Amazon Echo, Google 玩真钱的炸金花, and Apple 玩真钱的炸金花pod were revealed to secretly record user voices. In November, two researchers successfully hacked Amazon's latest smart speakers.
In the context of the ever-expanding smart speaker market, security issues continue to surface. What is the best of convenience and privacy?
Potential information security issues with smart speakers
The development history of domestic smart speakers is only two years, and Amazon launched the first speaker product Echo back in 2014, equipped with Alexa voice recognition system. In 2016, Google launched the Assistant ’s Google 玩真钱的炸金花. Since then, Alexa and Assistant have once been the object of ridicule by American talk show hosts, and artificial intelligence has become a “network celebrity.”
There is an old saying in China that "there is more right and wrong", which is used to describe the international celebrities Alexa and Assistant.
In April 2019, Bloomberg revealed that Amazon has hired a large number of employees to "monitor" smart speaker devices in real time. Even if the product's voice assistant is not activated, the user's private conversations will still be recorded by the device. And Google has also been pointed out that such a problem exists.
At present, intelligent speech products are far from being able to think and understand like humans, and the development of speech recognition requires a large amount of preliminary data to be imported for the construction of a "thinking bank".
The reporter learned that the speech recognition development company will manually perform a large amount of voice input, from content, intonation, and language to cover a wide range of content.
Obviously, the "database" used in the earlier development is far from enough in human brain capacity. Therefore, the product side will only hire employees. When users use the product, they will expand the "voice database" by collecting user conversations to optimize the product.
As soon as the 玩真钱的炸金花 came out, the incident was rapidly fermenting, and major technology companies have come forward to say that they have improved their functions.
Currently, Amazon Alexa has added a privacy option that allows users to turn off the ability to share voice recordings through this option, while Google has completely suspended manual transcription of Assistant audio.
This incident prompted the product side to improve its functions, but the "attacks" from third parties never stopped the "trial" of Amazon Echo.
How do smart products protect user privacy?
In 2018, the Tencent Blade team hacked the Amazon Echo in just 26 seconds.
In November this year, at the Pwn2Own hacking contest, two security researchers successfully invaded the latest Amazon screened speaker, Amazon Echo Show 5, and won a prize of $ 60,000.
Why are smart products that bring convenience to our lives so easily broken?
Zou Debao, a senior analyst of CCID Consulting's Electronic Information Industry Research Center, pointed out in an interview with a reporter from China Electronics News that before using smart speaker products, users must first enter personal information through the terminal on the Internet. Products in use will record people The sound is converted to a radio wave upload platform for processing, and anyone who can overcome the platform can capture the recorded sound.
As far as the domestic situation is concerned, the monitoring of smart products on the overall system top-level design is not enough, and there is no reasonable authoritative monitoring system for control. Security levels such as setting methods and propagation channels are not enough, which is a common problem faced by all product parties.
Zou Debao believes that the current re-engineering and remodeling of smart terminal products need to be imported through typical cases in the early stage to conduct small-scale supervision and control of products. For example, the introduction of the blockchain mechanism, multiple security encryption, the level and number of firewalls must be high enough.
Globally, the product form of smart speakers is no longer confined to traditional speakers, and "screen" speakers have gradually become the new darling of the market. The camera mounted on the speaker lifts the voice recognition to the level of image recognition, so from the perspective of functional design, how to protect the privacy of users?
Take the smart screen TV products launched by Huawei as an example. The camera on the screen has a lifting function. Users can choose to turn off if they do not need it.
From the perspective of hardware, Zhan Huayi, the chief scientist of Changhong AI Lab, proposed another possibility in an interview with a reporter from China Electronics News. You can use sensors to try to implement some functions of the camera. The sensor can sense the outlines of people and objects, but does not reflect clear portraits. Such a design can reduce the risk of user privacy leakage to a certain extent.
"The problem is not a negative situation. All emerging products must go through a long period of trial and error, and finally the relatively optimal solution is selected." Zou Debao told reporters.