Rage Against The Machine
Following on from my post two days ago (in which I wondered whether the scope of confidentiality could usefully be reduced in a translation context), the question is, given a great deal of the information translators receive is deemed confidential, does using Google Translate (GT) equate to disclosure of confidential info?
On disclosure, in an earlier discussion, I adapted an old Buddhist question – if I read out my “confidential” source text in the forest, but no-one hears it, is that disclosure….?
Black’s Law Dictionary was quoted back at me. “Disclose. To bring into view by uncovering: to expose, to make known; to lay bare; to reveal to knowledge; to free from secrecy or ignorance, or make known.”
Interesting stuff. I would say some of those descriptions imply, arguably even require, the involvement of another party – “bring into view” requires someone to perceive it; to “make known” requires another person to know the information subsequently. Some of the definitions don’t require another party, e.g. to expose or lay bare, but it could easily be argued that those that do not require another party’s involvement or presence might apply equally well to merely opening a document on a computer. And the Lord knows this line of work is sometimes painful enough as it is, without double clicking being deemed disclosure. Such a broad definition clearly cannot apply to translators – as long as no-one else can see the screen where the document is displayed. Like confidentiality, “disclosure” probably needs to have a more restricted meaning for translators.
Not only that, but it seems highly likely to me that, for the disclosure of confidential information to cause the detriment that I contend is a defining feature of the truly confidential, a human has to be the recipient, in order to a) understand the text and b) act on it such that detriment is caused. (I accept that other kinds of disclosure may not require a human presence, e.g. “disclosing” data to an automated stock trading system.)
On to the matter at hand. While Google clearly has legal personality (probably several, judging from its “don’t be evil” motto combined with its assiduous policy of tax avoidance), is a text passed through GT “disclosed” in any meaningful sense if a natural person is not aware of it?
Your decision on that can only depend on what you know, or think you know, about GT’s internal workings. From forum postings, some people evidently believe that Google don’t keep the source, it is merely transient or ephemeral within the GT machine, and therefore it cannot meaningfully be disclosed. Some think Google keep it or might keep it, and therefore it is at least potentially open to disclosure.
What the GT terms of service currently say is:
“When you upload or otherwise submit content to our Services, you give Google (and those we work with) a worldwide license to use, host, store, reproduce, modify, create derivative works (such as those resulting from translations, adaptations or other changes we make so that your content works better with our Services), communicate, publish, publicly perform, publicly display and distribute such content.”
That does not sound massively confidential to me.
Hence, as far as I am concerned, using GT constitutes disclosure, or perhaps more accurately, using GT runs an unacceptable and uncontrolled risk of disclosure. I doubt many humans actually see the content shoved through GT. But they most definitely could. And who knows what the future holds?
My guess is that those using GT:
– only use it for information that they are told is not confidential or is in the public domain
– only use it for information they think is not confidential or is going to be in the public domain soon
– do not view using GT as disclosure
– believe the risk of being discovered to have breached an NDA is small.
– a combination of the above.
But how can we reconcile today’s reality, in which some clients think everything needs to be handled confidentially whether or not disclosure would actually be to their detriment, there is uncertainty over what constitutes disclosure in a translation context, and software providers from all sides are exhorting us to release everything we have into the cloud?