Glencora Borradaile

         Associate Professor & College of Engineering Dean's Professor, School of Electrical Engineering and Computer Science, Oregon State University

November 27, 2017

IBM and the Holocaust — why wasn’t this on my radar?

Filed under: Silent Glen Speaks @ 7:17 pm
Tags: ,

I have spent the last couple of years teaching folks about surveillance and what they can do about it, primarily in partnership with the Civil Liberties Defense Center and through a interdisciplinary course on surveillance and social movements.  In reading about surveillance, I had come across, in multiple places, a statement along the lines of “IBM provided the punch-card technology that allowed the Third Reich to perform the census operations that were necessary to identify Jews,” with a citation to the book IBM and the Holocaust by Edwin Black.  I didn’t think too much about it.  After all, all modern surveillance is supported by technology, and I don’t necessarily place the blame on, for example, Intel for manufacturing chips that Hacking Group ultimately buys.  At worst, it seemed (from a single sentence) that IBM would be guilty of the kind of profiteering that many corporations operating in our capitalist world engage in.

Until I sat down to actually read IBM and the Holocaust; now I understand the significance of IBM’s involvement in WWII, and it can’t be summarized in a single sentence.  I will do my best to summarize in a few paragraphs (hopefully a little more succinctly than the book’s Wikipedia article).

Summary of IBM and the Holocaust

First, it is important to understand what punch card technology was like during the early 20th century.  The punch card tabulating machines performed basic statistical, sorting and selecting operations.  They were programmable only by experts and the machines themselves were not bought but leased from IBM (or other companies, although IBM had an effective monopoly on the technology).  In addition, the punch cards themselves were manufactured and supplied only by IBM for IBM machines, and they were manufactured for specific uses.  For example, punch cards would be printed to support the scheduling tasks of a specific railway system or the census of a particular region. The design of the punch card for a specific task was done by IBM employees in cooperation with the leasing group.  IBM directly trained the users of the punch cards and tabulating machines.  That is, IBM had intimate knowledge of the specific uses of their technology in the field.

Second, we need to understand who at IBM would know what their technology was used for.  Most of the tabulating machines used by Nazi Germany were leased by the IBM subsidiary in Germany, called Dehomag, which was 90% owned by IBM. Until Dehomag built its own factory in Germany (after Hitler came to power, and after census operations that would identify Jews were announced), the machines were imported from the US.  Likewise, due to supply shortages in Germany, punch cards were printed outside of Nazi-controlled areas well into WWII. The main reason, in my opinion, for the length of Black’s book is to provide sufficient evidence that Dehomag wasn’t a rogue subsidiary, but instead was heavily micromanaged by Watson (the CEO of IBM at the time).  Black makes it clear that Watson and many other executives at all levels of operations would have known specific uses of their machines.  Watson and other IBM executives made frequent trips to Germany until 1941.  Watson met with Hitler in 1937 and received a medal of honor from the Third Reich.  That is, IBM executives at all levels and internationally would be privy to the specific uses of their technology by the Third Reich.

Third, we need to understand what exactly IBM technology was being used for.  This is possibly the most disturbing part, especially in understanding how much IBM as a whole would have known about their uses.

I already indicated that punch cards and tabulating machines were used in census operations.  This is why IBM and the Holocaust comes up in surveillance literature — census is considered the most basic form of mass surveillance.  Census operations were called by the Third Reich in every territory it invaded.  Nazis were obsessed with identifying all Jews, including “racial Jews”, so they included not only religion on the census forms, but also the religion of ones parents and grandparents.  They would cross-check this with marriage forms and baptismal data held at churches.  Tabulating machines were used to understand how many “full”, “half” and “quarter” Jews lived in any district.  They also helped the Third Reich understand that as they imposed harsh regulations on Jews (e.g. Nuremburg Laws), many Jews would leave to neighboring regions, where the Third Reich would have to confront them again as they conquered neighboring territories, leading to the Nazi’s “Final Solution”: extermination.

Tabulating machines were housed in railway stations to help in the scheduling of trains and keeping track of the location of train cars.  Trains were used to transport Jews and other “undesirables” to and between concentration camps.  The sheer numbers of people being transported (in addition to war supplies) at the time required would not be supported by the 2 week delay in locating train cars that was typical in non-punch-card scheduling systems used at the time.

Finally, tabulating machines were housed at the concentration camps.  Each prisoner had a card that detailed their health, skills and location as prisoners of good health were transported according to labor needs.  Finally, the card also indicated the way the prisoner died: by natural causes (which would include being worked or tortured to death), execution, suicide, and special treatment (including gas chamber).

That is, IBM technology was used to support all aspects of the Holocaust, from identifying and transporting Jews, to managing their populations at concentration camps and recording their executions. Recall how much of this, by design of the punch cards and tabulating machines, IBM would know.

I want to include one story from IBM and the Holocaust which I think is important.  In all the Nazi-conquered territories, census operations made the rounding up and extermination of Jews possible.  However, they weren’t as successful in France, where an estimated 25% of Jews died.  For comparison, an estimated 75% of Jews in Holland died.  In France, the census operations were eventually completed by René Carmille.  However, Carmille sabotaged the operation, preventing any information about religion from being punched into cards.  He also used the census information to mobilize the French Resistance.  He was eventually captured and taken to Dachau where he died of exhaustion.

Why wasn’t this on my radar until now?

IBM and the Holocaust came out in 2001.  I started my graduate education in computer science in 2002.  It took me until possibly last year to even hear a hint of it.  I think this is an utter failure, utter lack of ethics education.  I was never formally taught any ethics.  I have to say that I was barely ever informally taught any ethics either.  Any ethical considerations I picked up during graduate school, looking back now, were not necessarily sound.  I know that I was just one student, but never have I even heard a colleague make mention of IBM and the Holocaust.  We should be talking about cases like IBM and the Holocaust with our students and with ourselves.  A majority of our students will go on to work for companies just like IBM.  And if they aren’t taught that tragedies like the Holocaust happen because everyone was just doing their job, we are liable for the continued abuse of computer science.

Print Friendly, PDF & Email


  1.   Jay McCarthy — November 28, 2017 @ 6:48 pm    

    You’ve convinced me! I’ll be talking about it with my students once I buy and read the book.

  2.   Alex Slivkins — December 5, 2017 @ 4:48 pm    

    Thanks for the post! I knew of the IBM involvement in general terms — same as you did — but not the specifics. Not too surprised, though.

  3.   William Gasarch — December 7, 2017 @ 5:06 pm    

    1) For putting ethics (or history or social impact or other non-tech issues) into the things-one-should-learn category there are two approaches:
    a) Have a required course on these matters. One problem is- who would teach it. Another problem is it get unconnected to the “real courses”- just one more requirement, like having to take calculus even though its rarely used.
    b) Having this kind of material as part of many courses. This will depend on the willingness of the faculty to do so. Might be a hard sell and some faculty might claim its not quite part of their course. And some may be right.

    2) CS has both an advantage and a disadvantage here.
    a) CS has still not figured out `every ugrad must know X’ hence we can be flexible in moving things around.
    b) CS seems to already be a 4.5 or 5 year major.

    3) Should IBM currently be held accountable for what they did in the 1940’s?
    I think not. I’d be more curious what companies NOW are doing that we should be aware of and, more importantly, what students should be aware of.

    Time to prove I’m not a robot. I think in the past this blog believes me, but
    Lipton-Regan seems to always think I’m a robot. Oh well.

    •   Glencora Borradaile — December 7, 2017 @ 5:38 pm    

      I do like the idea of having ethics, history, social context as part of many courses. But as you say, that takes a lot of buy in from a lot of faculty.

      As for IBM and responsibility, Black has an afterword about IBM’s refusal to help in the research of his book and there denial of any knowledge. While it might not be appropriate to hold the current IBM responsible (although, I personally may argue otherwise — after all, who does one ever seek reparations from?), I think most would agree that the right and good thing to do would to be to cooperate with the research into their past and apologize for their misdoings.

RSS feed for comments on this post.

Sorry, the comment form is closed at this time.

© 2017 Glencora Borradaile   Powered by WordPress MU    Hosted by