Responses to Comments on Huawei CV Paper

I make it a pretty strict rule to not read comments, positive or negative, about my work.  I have some wonderful friends or virtual commentators that have asked insightful questions that deserve some response. Given the number of comments and vociferousness this paper has elicited, I have decided to respond to some of the comments.

  1. The paper is not an academic paper. No, and I never said it was. It was never intended or designed as a journal article type paper for many reasons. The purpose of the paper was to provide information into the public domain that did not exist before in a concise and readable form for everyone from politicians in different countries to citizens hearing about the issue. The focus was on describing the data the specific profiles. That is it. It is not a good academic paper because it was never intended to be one.
  2. It is not an exhaustive study only taking a small sample of the CVs. Yes, I know I actually said that in the paper. This shouldn’t come as any surprise. There are two data or study specific reasons this choice was made. First, as I stated in the paper (ßhint hint) the data is messy and not easily given to general statistical analysis. This is being addressed to benefit future papers but given the computing capabilities needed and technical issues involved, this is not a snap your fingers and things are ready. Second, to actually analyze a CV one by one is a very time intensive process for a study like this. Just as one example, we needed to confirm through as many channels as possible that a CV of interest actually belonged to an actual individual. We needed to cross check information to business unit or military unit for instance. This is one instance where statistics will help provide some top line information but a lot of this work to confirm information about individuals is incredibly time consuming. To do this for any significant mass of individuals absent a small army of research interns is just not feasible. We will be following up on this and expanding on this work.
  3. So why didn’t you take more time and do a more exhaustive study? Public policy considerations.  In an ideal world, we would take 6-12 months and turn out an in depth and comprehensive study. The reality is that countries are making crucial decisions right now involving Huawei. The decision was made to proceed releasing this in the recognition that while the research would benefit from more time, the trade off of making this data available to country decision makers and the analysis outweighed the benefit of producing a more comprehensive study of the CVs. This data has been provided to some country governments already and will be provided to other countries looking to conduct their own analysis.  While I am sympathetic to the academic argument, in this case, I fully stand by the decision to proceed to market faster and make no apologies.
  4. The data and profiles are not replicable so we do not know how accurate they are. This decision was made fully knowing some would complain. Here is the logic. First, China has actually killed spies at the work place according to news reports. While it is unlikely they would kill one of their own, severe punishment of some kind of talking so openly about effectively classified or secret material would not be unlikely. We made this decision out of an abundance of caution to the targets. Second, while the decision to anonymize data was taken quite a while ago, a recent dispute between journalists about anonymizing subjects in China pieces clinched it that no information would be made public about these targets.  It is somewhat ironic that a week ago China watchers were complaining about subject data being made public are now complaining about data being anonymized. Third, as I stressed in the paper (ß hint hint) methods and data would be shared on a case by case by basis with trusted researchers. There is no intent to conceal or hide information from trusted people that guarantee to keep the data confidential.  This data and research has been shared with trusted researchers, governments, and politicians of all variety. We are more than willing to answer questions privately.
  5. Lots of companies, telecommunication firms probably disproportionately, around the world hire ex-military so Huawei is no different and the analysis provides nothing new. This is inaccurate for an important reasons. There is a clear line drawn between let us say the random chance of association and the specific work of the individual.  Assume a large company is recruiting at a job fair with other companies. We would expect through pure random chance that some of their 100 new hires would be ex-military. I do not allege this is a problem. In fact, this paper could have been written without this dataset. What we are alleging is that certain individuals covered in the paper and other individuals not profiled in the paper, went far beyond normal manufacturing, electronics, and or communications expertise for an employer. Rather, as I write in the paper (ß hint hint) I am alleging that certain individuals acted as intelligence assets for the Chinese Ministry of State Security (MSS) and the People’s Liberation Army Strategic Support Force based upon their own words describing the work they performed. I am not alleging that simply hiring ex-military makes an espionage organization. I am alleging that when an individual states on their resume that they are the MSS representative and engaged in behavior like information interception, that is work beyond normal corporate work but under the guise of the state. I have specific granular information of people effectively testifying to what they did and am doing nothing more than urge readers to consider the words of employees saying what they did.
  6. So what is the bigger issue about why this matters? Huawei lies.  They have been caught lying so much all the time they should have no credibility.  I have proven through Chinese corporate records that Huawei is not a private company, something Huawei has admitted in a conference call with press and for which Huawei produced a transcript. I have no shown not just that they have a close link with the PLA, but rather they have specific employees stating they are representatives of the state acting on behalf of the state while working for Huawei.  Again, these are their words.  The only thing I am trying to do is to put this information in the public domain.
  7. Huawei is no different as the US collects lots of data also. True and I sincerely hope there is greater focus in the US on greater legal restraints on the governments access to data, firms ability to share it, and how it is used. However, to equate US surveillance or Facebook data gathering with the level of intrusion and fear of reprisals faced from China reveals nothing more than towering ignorance.

We are already starting to plot out the next reports and reaching out to specific researchers with unique expertise to help us better analyze this data and put it in the public domain. There is a lot more to come.