Thin… brown line

C’est magnifique, mais ce n’est pas la guerre; c’est de la folie.
Pierre Bosquet

Light BrigadeThe controversial first episode of the first season of the much-acclaimed American television series The Newsroom, aired on June 24, 2012 in the United States, opens with the main character, a newscast anchor, participating in a debate and giving a controversial speech on America’s recent decline as a nation.

The controversy was due to the unpleasantness of the anchor’s arguments, although or maybe just because they are unquestionable facts and figures, so much so that, less than three years later, The Washington Post spread the news of the U.S. having the highest incarceration rate in the world — nearly 25% of the world’s inmates versus 5% of the world’s population — and, thus, more jails than colleges.

However, although unquestionable, some facts and figures are hard to accept and digest. Jeremy Rifkin’s latest book, for example, contains many of such stodgy facts and figures — even though they are no news — having been analyzed, reported and discussed a decade earlier by Robert Fogel in The Escape from Hunger and Premature Death, 1700-2100 and more recently by Erik Brynjolfsson and Andrew McAfee in The Second Machine Age.

Starting from its title, though, Jeremy Rifkin’s book addresses a very specific economic issue, which has also affected translation: the steady decrease of marginal cost, especially of immaterial products.

To use The Newsroom anchor’s words, “First step in solving any problem is recognizing there is one.”

Productivity is a measure of productive efficiency calculated as the ratio of what is produced to what is required to produce it. If the cost of producing an additional item (marginal cost) were nearly zero, that would be the optimum level of productivity. Technology boosts productivity. The new frontier that is being explored is Big Data technology, which will further increase productivity, and dramatically reduce the marginal cost of producing and delivering a full range of goods and services to near zero across the entire economy. In translation, the first application of this technology are free online machine translation engines, who in fact disrupted the industry.

The attempts to belittle this disruption are countless, now also from outsiders like Matthew Blake, Senior Vice President at Omnivere, an eDiscovery service firm, who recently wrote an article for Breaking Media’s legal news and insights website Above the Law, to acknowledge that translating foreign language documents during eDiscovery can be expensive and time-consuming, and that today’s law firms increasingly gravitate towards machine translation to speed up this process and reduce its costs.

In his article, though, Matthew Blake warns legal professionals to be careful when using any translation service that does not specifically protect data from any use not directly authorized by the client as accuracy and defensibility continue to surface, and suggests implementing machine translation behind a secure firewall operating under a terms of service agreement tailored to the legal industry as a most secure way to protect client confidentiality.

Jost Zetzsche was one of the first to rush in recommending Matthew Blake’s article in a tweet as a legal view on the topic, but soon after Smartling’s CEO Jack Welde made a point by signaling at the ‘Sponsored Content’ header above the title of the article.

As Jack Welde rightly observed, Matthew Blake’s article is, in fact, more an advert than a legal opinion, intentionally misleading readers by referencing Google’s general terms of service only, while the terms of service for Google Translation API are very clear as to data confidentiality and ownership.

In his article, Matthew Blake also adds up that even legal professionals “won’t have a method of determining translation accuracy across the body of documents for a specific case,” thus making the case, possibly unknowingly, for information asymmetry, which is a typical issue in the translation market.

In fact, Jack Welde contends that the most compelling argument against using Google Translate for eDiscovery is the quality of translation, even though he seems to forget what eDiscovery is all about — locating and identifying data that are potentially relevant for a legal hold — and even Google Translate could be suitable to read the meaning of a document that could otherwise remain just legal gibberish in a foreign, incomprehensible language.

Data confidentiality and ownership are, in fact, sensitive topics, and many MT-scared neo luddites are leveraging them as clubs against the failures of translation technology, as if security were not a concern before the advent of machine translation, the Internet, and the cloud.

These neo luddites are actually trying to curb a raging river with sandbags, emptying the hull of a sinking liner from water with a glass, charging an overwhelming armada with a light brigade. There is no camp to defend and no thin red line to keep. What is worth in fighting a senseless lost war, sponsoring a lost cause, or championing an inane backward march?

Just as for cloud computing, machine translation is not for everyone and it is not for everything. Just because it is here (to stay, actually) and offers some benefits this does not mean it should be used anyway.

The same goes for data ownership in general. Data is not safer on one’s own computer (as the only really secure computer is a dead computer). On the contrary, it is most probably safer on someone else’s if this is a dedicated service company. In fact, if a computer can be owned, and thus manipulated ad lib, when data is manipulated with third-party applications it should not be forgotten that, in respect to software, we buy is a license to use, not the product itself, and this applies as to desktop as to web applications.

So, if data to be stored on a provider’s cloud can be encrypted and protected from reading — not from stealing or destruction — the same data cannot be encrypted if it is to be processed. Therefore, even though most end clients do not care about the details of their vendors’ processes and tools, unless they have a certified quality system in place, signing a terms and conditions agreement detailing data handling procedures is always advisable.

In other words, as it has been happening at least for three decades — first with the appearance of the personal computers, and then with CAT tools, the Internet, and with now machine translation and the cloud — there is no reason to be scared of technology. We should learn to coexist with it and exploit it, rather than fear it and (uselessly) fight it.

A real professional is capable to tell whether and when to use cloud computing and/or machine translation.

Anyway, to quote Jeremy Rifkin, connecting everyone and everything is bringing humans out of the age of privacy and into the era of transparency. And it is curious to see how the same people eagerly sharing every thought of theirs, every moment of their lives with the rest of the world via Facebook, Twitter, YouTube, Instagram, and the like, are so worried about privacy, data confidentiality and ownership only when it comes to a technology they feel as a threat for their status.

Well, the news is that 75 percent of the leaders of global companies surveyed by The Economist are exploring or using the IoT in their businesses to some extent. Maybe confidentiality is not a real issue when data gets to a translator…

Evidently, the anxiety with technology, and especially machine translation, makes even the bravest ones get lost in a glass of water.

Author: Luigi Muzii

Luigi Muzii