Automation of Telephone Call Monitoring in China

[ After writing this blog article, I got an interesting tweet from China about phone monitoring. 重要对话用 @telegram 端对端加密,电话短信早就被监控着了。” For important conversations use Telegram Messanger ( https://telegram.org/ )with its end-to-end encyption. Telephones and instant messaging have been being monitored for quite a while.”]
 The Economist on January 27 published the article “In China, consumers are becoming more anxious about data privacy — will this impede government snooping” about how Chinese are getting more concerned about data privacy.  China has some of the most active e-commerce websites in the world with much of the activity coming from the smartphones in the hands of hundreds of millions of Chinese.

Consumers in China have good cause to worry. Data collected through one medium can often end up in another. A man who talked on his mobile phone one day about picking strawberries said that when he used his phone the next day to open Toutiao, a news aggregator driven by artificial intelligence (AI), his news was all about strawberries. His post on the experience went viral in January. Toutiao denied it was snooping but conceded, blandly, that the story revealed a growing public “awareness of privacy”.

In the United States, at least until very recently, most people have generally been much more concerned about government snooping about their phone calls and internet data than with corporate snooping.  When they sign up to use many free smartphone applications, people often sign off to access to many functions (microphone, camera) to the app sponsor without considering how much of their privacy they are giving away to a corporation that might well sell it on to some other data aggregator.

I wonder how that works in China?

Telephone Monitoring and Me in China

I have always been curious about telephone monitoring since I was the object of telephone monitoring during the ten years I worked in China as a U.S. diplomat.  For the first month while I was at U.S. Embassy Beijing in 1996, my telephone had a funny humming noise.  I wondered whether that was because the monitoring people had bad equipment or because they thought they could intimidate me that way.  The funny hum went away after a month.  I supposed that the monitors must have decided that I was actually just a boring diplomat instead of somebody more exciting like a spy.

 

After that I only heard the funny home on my cellphone when I was travelling — but never while I was home in Beijing.  I always wondered — could the monitoring equipment be that bad or did they want to remind me that my conversations were being recorded. From an intelligence collection perspective, reminding me with that helpful hum that my phone calls were being recorded wouldn’t be a good idea.  One of the things I learned during my career is not to be too quick to think people are out to get you when incompetence is often a perfectly plausible explanation.  I never figured out just what was going on.

There were huts on top of all the buildings in the Tayuan Diplomatic Compound in Beijing where we lived. One time on of my colleagues told me how his five year old son was walking on the stairway when he saw that the door to the stairway leading to the roof hut was open.  He later told his father that he walked up the stairway and saw a man inside “with all kinds of computers and stuff”.  Perhaps that was someone changing tapes or adjusting the recordings.  Next to the compound was a five-story telephone exchange building — I imagined that there must be many people in there listening to phone calls in many foreign languages.

Ten years later, when I was working at the U.S. Consulate General in Chengdu,  I was walking with my friend, the now-deceased Chengdu writer Yin Shuping, past a very big telephone exchange building in the county seat city where he lived.   I said I was surprised at how large the building was.  Yin answered, “That’s because they need a lot of room for all the people listening in on telephone calls. Every county seat in China has a big telephone exchange building for that.”

I have always been astonished at the size of China’s domestic security workforce. For example, I remember when President Clinton visited China in 1998, his motorcade drove past my apartment at the Tayuan Diplomatic Compound there were miles and miles of  Chinese plainclothesmen every three feet or so.  They made it clear that they didn’t want to be in my photos either. So I figured they must be real plainclothesmen and not just some random people helping out.

In China the size of the police response to say a few demonstrators seemed wildly disproportionate to the number of protesters. Maybe they were worried that a “single spark can light a prairie fire” like Chairman Mao used to say. China’s domestic security spending exceeds China’s total military spending.  In the U.S., total annual spending on policing is over USD 100 billion while total military spending (2015) was about USD 600 billion. Perhaps I shouldn’t be so astonished since the U.S. military is so large and since I live here I consciously and unconsciously take the U.S. experience as a reference point.  I wonder what the ratio of police to military spending is in other countries.

Voice Recognition as a Tool for the Authorities in China

I wonder if this means there is widespread use of voice recognition technology on phone calls in China?  Perhaps voice recognition technology is lightening the phone call monitoring load for the Party.  Monitoring conversations in the various dialects of Mandarin and indeed the many different languages — Shanghaiese, Cantonese, Fuhouhua, Chaozhouhua, Minnanyu etc. — in the Chinese language family must be quite a challenge!

Looking around online, I saw a website offering a Chinese language voice recognition product that claimed to be suitable for generating transcripts of courtroom proceedings — actually saying it is for “intelligent courtrooms”.  One would hope that all courtrooms are intelligent courtrooms so perhaps a better translation would be cyber-augmented or something.   I hope the automatic court transcript generating system is reliable!

From https://cloud.tencent.com/product/asr

语音识别ASR

语音识别(Automatic Speech Recognition)服务,应用业界最先进的深度学习算法,具备出色的语音转文字、关键词检索、静音检测、语速检测、情绪识别能力。全面满足电话录音质检、实时语音输入、直播字幕及审核等多种场景下的语音处理需求。
随着人民法律意识的提高,人民法院的案件审理数日益增加,庭审效率亟待提高,而庭审现场记录的百分百还原,一直以来都无法实现。通过在法庭现场应用语音识别,可以将法官、原告、被告等各角色的现场发言实时转写成文字笔录,并对录音及文本进行存储,减轻书记员工作负担的同时实现了庭审现场的百分百还原。专业法务领域定制化引擎,有效应对法律专业术语。支持自定义热词,提升人名、机构名、公司名等词汇识别准确率。个案定制话模型,满足物理、化学等专业领域司法案件的高准确率需求。私有化部署,保证人民法院的数据安全性。
Looking further I found a discussion of voice recognition technology in connection to the case cited in The Economist article above. The gist of the article: the Chinese corporate website belonging to a company in the Alibaba e-commerce group said, ‘No,  it isn’t true, that just isn’t possible with current technology. But the journalist asked a university prof expert on voice recognition technology Professor Xu Mingxing of Tsinghua University who said just the opposite “Current technology makes this automatic voice recognition of a phone conversation in this case entirely possible.”

Excerpt from the article  百度读取通讯录被告 今日头条陷“窃听风云”(Baidu Accused of Monitoring Record of Communications,  Jinri Toutiao Is Caught up in the “Bugging Cloud” Controversy)                    http://azcnews.org/20180107/%E7%99%BE%E5%BA%A6%E8%AF%BB%E5%8F%96%E9%80%9A%E8%AE%AF%E5%BD%95%E8%A2%AB%E5%91%8A-%E4%BB%8A%E6%97%A5%E5%A4%B4%E6%9D%A1%E9%99%B7%E7%AA%83%E5%90%AC%E9%A3%8E%E4%BA%91/

— “声音识别技术窃听用户隐私  (Using Voice Recognition Technology for Surreptitious Monitoring Invades User Privacy)
根据手机百度官方网站的资料,手机百度“是一款有7亿用户在使用的手机‘搜索+资讯’客户端”。无独有偶,几乎同一时间,用户数达6亿的手机软件“今日头条”APP也遭网友指控涉嫌透过手机麦克风窃听用户隐私,并推送相关新闻和广告。网民“互联网络人”在新浪微博帖文中说:“妻子元旦去摘草莓,一没在头条搜索草莓,二没在头条看草莓类资讯,隔天就收到了和草莓有关的推送文章。”这位网民表示,事后发现“今日头条”使用了手机麦克风的权限。有网民反映,在和朋友讨论西餐厅哪家好,但从未在其它APP或者浏览器上做任何搜索,“今日头条”会马上推送与西餐相关的广告和资讯。有网民说,她在微信上问朋友元旦去的温泉叫什么名字,然后“今日头条”马上推荐温泉酒店。还有网民表示:“我发现过很多次讲话的东西被推荐。”

除了“今日头条”的“窃听风云”,淘宝网也被认为是窃听“重灾区”。据科技新报网报导,近年来,不断有网民反映,与朋友聊天聊到的某类产品,很快会出现在淘宝推荐中,而之前自己从未在淘宝上搜索此类产品。

1月4日,“今日头条”官方回应,称从技术上看,目前声音信息的处理,远达不到通过麦克风去获取个人隐私的水平。但是有声纹处理专家却持不同看法。

据雷锋网1月5日报导,公众号“雷锋网”就该技术求证清华大学—得意音通声纹处理联合实验室副主任徐明星,他认为,从技术上讲,手机APP是完全能私自录制用户日常语音的。把语音数据传给语音识别引擎,语音数据就可以被记录、传输和保存。“以目前的自动语音识别水平,可以胜任面向广告推送的关键词抽取,然后经过文本自然语言理解的一些技术,就能掌握用户的一些个人需求,从而推送相应的广告。”

有网民表示,手机乱七八糟的程序很多,看手机设置里就知道。很多程序都会获取手机识别码,获取手机联系人,可以监听、拍照、录像,一些常用程序如微信更甚。所以,用手机通话,没有任何私密性可言。“手机就是一个安全部的窃听器。”

1月3日,阿里巴巴集团旗下的软件支付系统支付宝年度账单捆绑推广芝麻信用被曝光,支付宝年度账单的首页通过默认同意的方式,让部分用户在不知情的情况下“被同意”接受芝麻信用,个人的金融信息被搜集处理。芝麻信用随即发表致歉声明并修改页面。

I wonder how effective this automated monitoring is?  Does it just pick out and flag some key words or does it do get enough of a conversation to do more useful monitoring that can pick out from many thousands of telephone calls the several that are worth the time of a human operator?   Background noises and the different dialects and languages spoken by people on the phone make monitoring more difficult.  I imagine with the massive investments now seen in China for domestic political security there must be quite a bit of work underway to develop these systems.
With the parallel applications developed for market research in the shopping trends of hundreds of millions of Chinese customers, there must be many commercial applications that are finding domestic security applications with the authorities, both in China and in other countries.
Improved technology now means that China can now cut its telephone monitoring workforce or it can now monitor a much higher proportion of all phone calls.
Maybe the next time I go to China I should talk on the phone in Pig Latin. Maybe there is not a voice recognition app for that yet!
Advertisements

About 高大伟 David Cowhig

Worked 25 years as a US State Department Foreign Service Officer including ten years at US Embassy Beijing and US Consulate General Chengdu and four years as a China Analyst in the Bureau of Intelligence and Research. Before State I translated Japanese and Chinese scientific and technical books and articles into English freelance for six years. Before that I taught English at Tunghai University in Taiwan for three years. And before that I worked two summers on Norwegian farms, milking cows and feeding chickens.
This entry was posted in National Security 安全, Science, Technology and Academic 科技学术, Society 社会 and tagged , , , , , , . Bookmark the permalink.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

w

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.