Corrosion of confidentiality
Another flurry of comment of late on the evils of MT, especially Google Translate (GT). Plus there were, I think, a couple of conferences on the subject of MT this autumn, judging from my Twitter feed.
I found myself in something of a peculiar position in this regard. I don’t actually use MT myself out of concern for the general perception of the confidentiality issues it brings (it is, after all, easier to go with the flow and not enter into protracted discussions with clients) but I initially felt that I understood or at least empathised with those who do use it (on the understanding I’m talking here about its ability to save typing effort or generate ideas, not produce deliverable output).
But as ever, I started to wonder, this time on the nature of confidentiality…
I should say from the kick-off that I find some translators have a tendency to yell “confidentiality” to support their position on anything they don’t like or don’t want to do, and this reflex response raises my hackles. The sort of person who will redact the life out of a phrase they post on a terminology list, and then when asked for a bit of context or background (e.g. what sort of organisation has produced the document), will reply “sorry, confidential”, while seemingly merrily unaware that copying a handful of words from the phrase into a search engine will, surprisingly often, lead directly to the text being translated. Fie, a pox on your spurious confidentiality claims, say I.
Anyway….
Most translators have probably signed confidentiality agreements/NDAs and I have also translated a fair number (which hardly makes me unusual – not for that reason, anyway). They are all similar but never quite the same. The average message, if such a concept exists, appears to be “a thing is confidential, and therefore not to be disclosed, if we say it is, and we say everything we pass on to you is confidential”. There may or may not be a rider to the effect that you will be spared the lash if information is in the public domain already, or enters it without your being at fault, or if you disclose information to people who it turns out already knew it before you went around leaking information like a cracked colander.
Translation NDAs rarely, in my experience, go much further. I think I’ve seen just one that exhorted me to think carefully about what I post when asking for terminology help in public (something I stopped doing in any meaningful sense long ago anyway).
But what, then, is the point of confidentiality? There is undoubtedly stuff you would at some level prefer people not to know, unless they need to know. Is all of that stuff really “confidential” in the sense that it warrants sanctions to dissuade against disclosure?
Purely out of a sense of decorum, I’m not going to tell you colour of my pants. Purely out of some sense of it being none of your business (my belief that you have no need to know), I’m not going to tell you what I earned last year. And yet if I were 30 years younger, I could potentially adopt a dress code displaying my pants to all and sundry, and you’d know whether you wanted to or not. If I lived in Norway, my earnings would be on a website. Maybe those examples overlap a little with personal privacy, but I daresay I could find large numbers of suit-wearing, middle-aged Englishmen willing to describe their tax returns, if not their pants, in the same terms as the documents in the pulsating offices in which they toil, i.e. as “confidential”… but are they?
I suggest they are not. I’ve searched for an hour or two, and failed to find any legal definition of what any jurisdiction deems it reasonable to categorise as confidential, or in contrast what it deems unreasonable (for example in a case where, against a charge of disclosing confidential info, the defence was that the information, like the colour of my pants, not actually confidential at all.). The stance everywhere seems to be that if a party says it’s confidential, then it is. So I’ve had to think of my own definition of what I think the real scope of NDAs should perhaps be, in a translation context.
I say in a translation context, because it seems to me several factors are acting (and have been for some time) to drive us out of our isolation, with a corresponding impact on any all-encompassing notion of confidentiality. These include the development of cloud computing, SaaS, and online tools (e.g. Wordfast Anywhere), the current fad for “collaborative workspaces” and similar, the cumulative effect of the promotion of MT (“it’s good, really, you should try it, everyone else is, you’ll get left behind…”), especially online versions (e.g. Google Translate, naturally), and indeed the rise of general and open social networks as places where it is possible discuss terminology and seek help (in contrast to the previous situation where such things were restricted to translator websites, which, while there was still the risk of disclosing confidential info, you were at least disclosing it to people less likely to drop you right in it, on the do-as-you-would-be-done-by principle).
I have read comments suggesting that using any of these services may well be in breach of confidentiality, and one certainly does wonder (notwithstanding my earlier remarks about the propensity to yell “confidentiality” to object to anything and everything).
And that in turn leads me to wonder the following: if we cannot be guaranteed as to the confidentiality of such services (even for those services that claim confidentiality, such as Wordfast Anywhere, the truth is the situation is out of our control), would it be useful to redefine what “confidential” covers?
Two options seem to present themselves:
1. The aim of NDAs for translators ought to be only the prevention of the disclosure of information where such disclosure is or potentially could be demonstrably to the detriment in some way to the interests (financial, reputational, or any other) of the party which “owns” the information (itself a notion that is something of a legal minefield – I suggest for our purposes, it means the end client), this to include where the information is of demonstrable benefit to the interests of the receiving party or any potential receiving party which the information owner did not intend them to have at that point in time.
And leaping briefly back out of a translation context, the truth is, I can’t think of a way your knowing the colour of my pants or my earnings last year is of benefit to you, or works to my detriment. Other than, in a rather circular manner, harm to my reputation as a fellow to be trusted with confidential information, should I in contrast appear to be a man willing to blurt his income to the world and his wife or civil partner. Except, I would contend, I can think of no possible benefit to you or harm to me from your knowing either of those things. Therefore they are not confidential and therefore my reputation is intact. QED. Or so it would be hoped, anyway…. Hence there clearly is information that while not confidential (by my definition) nor yet in the public domain, could at least be revealed on a fairly broad “need to know” basis, rather than classified under a blanket confidentiality label as now.
In exchange for this restricted scope of the information content that is deemed confidential, the NDA should in addition potentially include the fact a particular document is being translated at all (case by case, where this fact meets the above definition), and/or the identities of the parties involved in the translation process (note that, unlike some, I do not have a list of satisfied clients on my website).
So, as an example of what I have in mind, taking a hypothetical consumer product sheet of some kind:
a) “the packaging is red” – not confidential; publicly available information.
b) “the packaging is red because it’s the Chairman’s wife’s favourite colour” – not, I contend, actually confidential. I may not need to know, admittedly. But I fail to see any harm in anyone knowing.
c) “the packaging is red because we tested lots of colours and this worked best for this product” – yup, no argument with the confidentiality of that, obviously
d) “the packaging is red [+ the fact this is being translated into Chinese] – confidential as an indication of future market intentions, f’rinstance.
And so, when projects are handed to us, we are told, for each one, whether it is to be treated as “confidential” or not under the above definition or one similar to it (I would naturally expect a good deal of our work would in fact then not be confidential at all).
2. Translator NDAs could acknowledge that such tools exist, that our working environments are changing, and make appropriate provision for them. Explicitly permit or prohibit all of them, or some of them – the difficulty is, of course, the constantly shifting landscape. And the fact that too many agencies are just looking to shift the work while exposing their own arses to the minimum risk.
That is perhaps a basic flaw in both suggestions, in that they require those occupying the space between translators and their end clients to actually give some thought to their processes and procedures.
I am not suggesting a free-for-all. I am certainly not suggesting that any translator (and definitely not me!) unilaterally decides to reinterpret any confidentiality agreements, in line with the above notions or any notions of their own devising. Any re-drawing of the boundaries (new NDAs) has to come from the information owners or agents acting on their behalf. It might be useful if those engaged in the sale and marketing of products and services that appear to require us to contravene confidentiality agreements gave the matter some proper thought. They could perhaps influence the decision makers, like they do when they’ve got units to shift. Who knows, it might increase sales, and all that.
Meanwhile, if the information owners and middlemen say everything is confidential, then it is. While that remains the situation, I’ll sign the NDAs and I’ll adhere to the NDAs, and in so doing, I’ll restrict the potential options open to me to provide a better service to clients. And I won’t be spending any money on any cloud-collaborative-anywhere-machine-as-a-service gubbins. So be it.
And since this is already much longer than I originally anticipated, I’ll save the bit about GT for another time.
I’d take a practical minimalist approach and suggest that confidentiality is only relevant when it is actionable, ie the client can only claim it where he can demonstrate a direct loss flowing from the breach. Such cases would be far and few between.
Many of the docs I have been asked to translate are actually public docs, ie court/tribunal decisions and notarial documents. The whole purpose, in fact of notarizing a document is to make it public in an official and recognised manner.
No expert but I would assume that from the moment you are emailed the doc to translate, if said email is not encrypted then a potential breach of confidentiality has already occurred.
In the circumstances I’d interpret my duty of confidentiality as that I will do my level best not to disclose this information, being aware that this may: a) conflict with my duty to give the best service possible, noting that clients often want a quick service, that’s were MT and emailing docs come in. And b) they do not impose unnecessary restrictions on MY freedom.
And as you say, who cares what colour your pants are anyway???
Thanks for respondong. I see your point, although I am not sure I would want to take it quite that far, if only because I’m not sure (end-)clients would. I think an ability to show discretion and treat sensitive information properly builds better trust. It might be impossible to demonstrate a loss from, e.g. disclosing that a company is looking into the possibility of expanding into country A without having yet decided yes or no, whereas I can see it is not ideal for their competitors to become aware of it. Detrimental, if you like 🙂
I’ve seen the point about email before. I almost included it. I think I agree in principle, despite being a gmail user (I do have my own domain etc. but for some reason rarely use it).