Posted by: Steve Vitek | October 19, 2012

The End of Human Translation Will Be Most Likely Postponed Again By a Few More Years (Or Decades or Centuries)

I read on a blog post from 2010 yesterday a modest prediction that within a few years, 90 percent of “LSPs” (Language Services Providers) would be out of business because they will be replaced by companies using customized machine translation systems. This prediction was made, surprise, surprise, by a guy who is selling customized machine translation systems.

First of all, I never use the term “LSPs” which was originally crafted in the nineties at some conference to include both freelance translators and translation agencies if I remember correctly what Chris Durban, a well known financial translator based in Paris, once said in a comment on my blog, because the word “LSPs”, which is now a synonym for the more accurate term “translation agencies”. “LSP” is these days used to distinguish translation agencies from freelance translators, although many freelance translators are at the same time also translation agencies, present party included.

I think the term was designed primarily to avoid the use of the more accurate term “translation agency” and somehow make it seem that it is the agency rather than the translator who is the real provider of translations.

Since the real providers of translations are obviously translators, not the agencies, I have suggested that the term “LSRs” (Language Services Resellers) would be perhaps more appropriate, although the term “translation agency” is obviously the best because nobody outside of the translation industry understands the abbreviation “LSP” anyway although it has been thrown around by agencies and translators alike for something like 20 years.

Secondly, this particular salesman of customized machine translation systems estimated in that particular post, which is now 2 years old, that there were about 10,000 “LSPs” in business at that point. I take it he meant translation agencies worldwide.

How the hell did he count them? If you look in your local Yellow Pages, the chances are that you will be able to count at least half a dozen “LSPs” right where you live, many more if you live in a metropolitan area, although most translation providers no longer even advertise in Yellow Pages. I certainly don’t. It is a waste of money because unlike in the 20th century, you are much better off these days if you have a well designed website.

I think that there must be more than a hundred thousands translation agencies on this planet, from Puget Sound to Punjab and from Beijing to Brusselles, some kind of big and some kind of small. Although probably at least 20 percent of them do go out of business every year, they die not because they have been replaced by customized machine translation systems, but because they fail to identify suitable clients and keep these clients happy.

It would be so nice if it were possible to replace human brain by customized software in language translation, wouldn’t it, especially since this is something that already happened in so many fields. Bank tellers became all but extinct as most of them were replaced by ATMs and software many years ago. So did phone operators, for example, as well as many mid level managers.

Our modern version of corporate capitalism loves machines that do the same job that humans used to do, except virtually for free. As human labor was first outsourced to wherever it was cheapest, every couple of decades it was moved to different countries, first to Mexico, then to India and China. Next decade, who knows … Vietnam and North Korea? I hear the labor cost is much cheaper there and best of all, workers in prisons there have to work for free! That’s even cheaper than child labor!! Booyah!!!

If you can replace dirt cheap labor by customized software, it’s cheaper still. It must be music to the ears of thrifty managers responsible for translations in large companies when they hear that they can save all this money that they have been spending on human translation by simply purchasing a customized machine translation system.

Are they buying? I don’t know. If the types of businesses that sell these machine translation systems are still in business although machine translation systems such as Google Translate (Mox calls it “Gurgle Translate!!!) or Microsoft Translator and many other systems are available for Internet for free, some companies must be buying them.

But were 20 percent of translators like you and me already replaced in the past two years by customized machine translation software? What do you think, gentle reader? Do you know a translator who has been replaced in this manner? If so, I hope you will leave a comment on my blog.

I personally think that because it is kind of really, really difficult to replace human brain such as yours or mine by software, no matter how carefully and craftily the software has been designed, the real question is how many businesses selling customized machine translation system will be still gracing this planet with their presence eight years from now.

After all, since Google Translate, Microsoft Translator, Babelfish and many other machine translation systems are available on the Internet for free, it would be hard to compete with them on price, and for machine translation systems, they work pretty well.

About these ads

Responses

  1. Well said. I just wish “translation agencies” would stop calling themselves that way since they do anything but translate, most of the time (and less and less so)! This is just as an abusive language as calling themselves “Language Service Providers”! These people are empty shells, intermediaries that freelance translators do not need for their survival, whereas those empty shells do need freelance translators for their survival! What’s shocking is that, since the end-customer pays them, they take the amount of money they need (plus a fat margin) and leave a few, insufficient, peanuts to the person who did most of the work! I strongly recommend to stop working via those empty – but greedy – shells and to contact end-customers directly! Amen.

    Like

  2. “I strongly recommend to stop working via those empty – but greedy – shells and to contact end-customers directly! Amen.”

    I see them more as a necessary and somewhat beneficial evil, such as taxes or marriage, especially since I often function as an agency myself although at heart I am really a translator.

    The Electoral College here in US on the other hand is in my opinion a completely unnecessary evil.

    Like

  3. “Everything we call `a necessary evil´ becomes more and more evil and at the same time more and more (un)necessary eventually.”

    I don’t remember where I read this quotable quote. Redears’ Digest, maybe?

    Like

    • BTW, I do work with agencies and I think what Kevin wrote in this blog post is perfectly all right. (Note: I “work with” agencies, not “for.”)

      http://www.translationtribulations.com/2012/09/we-are-wind-beneath-their-wings.html

      Like

    • Talking about evils, the bigger ones and smaller ones, it is also true that what one may perceive as the smaller evil is often in fact the more effective evil precisely because it is being perceived as the smaller evil.

      And the thing is, the more effective evil is by definition the bigger evil because it can do a lot of damage without people noticing it.

      I am obviously talking about the way people have been deciding for decades in this country whether to vote for Democrats or Republicans.

      Like

      • In Chinese, we say 兩害相權取其輕 to the decision on a bigger or a smaller evil.

        It means, “Two harms (兩害) compared to each other (相權), you would take (取) the lighter (輕) one (of them 其).”

        Like

  4. [...] I read on a blog post from 2010 yesterday a modest prediction that within a few years, 90 percent of “LSPs” (Language Services Providers) would be out of business because they will be r…  [...]

    Like

  5. It is true that machine translation systems cannot get sick, protest low fees, or feel tired or out of sorts (arguments which may be wielded at some time or other by its promoters) but on the other hand, human translators cannot be hacked, nor powered down, nor their algorythms corrupted… :)

    Like

  6. Human beings cannot be hacked (except with a sharp instrument), but they can be brainwashed, which works pretty much like hacking in cyberspace.

    Like

  7. [...] I read on a blog post from 2010 yesterday a modest prediction that within a few years, 90 percent of “LSPs” (Language Services Providers) would be out of business because they will be r…  [...]

    Like

  8. [...] I read on a blog post from 2010 yesterday a modest prediction that within a few years, 90 percent of “LSPs” (Language Services Providers) would be out of business because they will be r…  [...]

    Like

  9. “I personally think that because it is kind of really, really difficult to replace human brain such as yours or mine by software, no matter how carefully and craftily the software has been designed, the real question is how many businesses selling customized machine translation system will be still gracing this planet with their presence eight years from now.”

    I completely agree. Can somebody tell me if a machine translation system is able to go beyond the meaning of words to translate them accordingly, while taking into account other aspects, apart from the isolated meaning of words? Can a machine translation system really adapt the text to help the client get what he or she wants? The problem with MTS vendors ;) is that they don’t understand that we, translators, don’t translate words, but cultures, and this is something very different from what a machine translation system does.

    Like

  10. Thank you for your comment.

    I have been translating a complicated insurance contract today. I thought that it would help to use machine translation, so I translated it with both Google Translate and Microsoft Translator, but the results were pretty horrible in both cases.

    We heard that “machine translation is not quite there yet” fifty years ago, and people will be hearing exactly the same sentence 50 years from now.

    Like

  11. I think that another problem is when technology is in the hands of translation agencies because, in most cases, they force their translators to use it in order to pay them lower rates (as it is happening with Post-editing), while they continue offering their services to the end clients at the same rates. This, of course, doesn’t do anything but increase what a translation agency earns.

    We should never forget that technology must always be under our control and we must be the only ones to decide when to use it. As Rose Newell once said: “We should use technology to assist us. We should not allow technology to assist others in using us.”

    Like

  12. Well said.

    But I think that mostly large agencies subscribe to this predatory business model.

    Small and tiny outfits are in my experience much easier to work with because they are interested in establishing a good relationship with their translators.

    Like

  13. This is the comment that clarified the GTS version :

    I have been saying that there is a coming consolidation in LSP land. Let me explain more clearly what I meant.

    There are perhaps as many as 20,000 “LSPs” according to CSA – many who only provide brokering services with freelancers of unknown quality and provide the most rudimentary project management services if any. These are the people who will likely be marginalized and removed from the supply chain. In the current market situation it is often hard to tell the low quality from the high quality producers because they all seem to say the same things, but as process and production automation improves it will be easier to identify opportunistic brokers who push down prices and deliver crap and thus hurt the whole eco-system and profession.

    I believe that SLVs who produce consistent high quality and are known to be fair to their translators are likely to be highly desirable partners for both the MLVs and the Global Enterprises who choose to bypass MLVs and go straight to the quality production players. Collaboration and process management technology makes it much more possible to bypass MLVs who add little value but charge “project management fees” for handing off the work to SLVs. The best MLVs will need to do more than project management, and play a meaningful value adding role in initiatives like developing MT engines, community/crowd management, increasing process automation, JIT translation, user generated content translation management and handle the many ongoing flowing streams of content that drive global business. There are thousands of LSPs who can do basic project management but very few who can really provide value in MT or community engagement and other quality management initiatives.

    While I do predict that many LSPs will be marginalized in this coming world, I also predict that the value of the best translators will continue to rise as global enterprises realize that they are key to making high value content multilingual in a cost-effective and timely way but yet maintain target quality levels. It is possible that some or many of these “best translators” will learn how to make MT work for specific purposes with available resources. I also predict that the best translation technology will come from those ventures where there is a respectful, cooperative and collaborative partnership between technology developers and skilled language translation professionals. There is a bigger role for skilled language professionals in MT than post-editing. Hopefully this starts getting clearer over the coming years

    Like

  14. Thank you for your comment.

    I am not even sure what all of these abbreviations that you use mean, the only two that I recognize are LSP and MT.

    It might be a good idea to explain next time abbreviations to people who don’t necessarily live in your world.

    Here is what I think:

    I think that it is predominantly the large agencies who usually deliver low quality. Since the profit margin is extremely important to their business model, they can’t afford to pay higher rates, which is why they must often work with beginning translators. Once the translators learn how things work, they usually leave the low rate leaders for greener pastures.

    Large agencies are probably also likely to adopt and integrate edited machine translation in their business model if that can further increase their profit margin, which is in turn likely to result in further deterioration of the quality.

    Smaller translation agencies and individual translators specializing in a certain market segment are better positioned to deliver better quality, provided that they know what they are doing and that they stay away from edited machine translations. They will still use machine translation as an additional resource, but not as a translation that can be simply edited and sold.

    That makes them better able to survive the current economic climate, provided that they mostly work for small and medium-size companies and stay away from large corporations who mostly care only about the bottom line.

    I am not saying that there is no place in the translation market for people like you who are trying to sell a different translation model based on machine translation, and I wish you luck.

    But to say as you did in your post 2 years ago that in 10 years, there will be hardly any LSPs left who are not using your approach is really arrogant and transparently self-serving.

    Which is what prompted me to write my post.

    Like

  15. Steve

    “But to say as you did in your post 2 years ago that in 10 years, there will be hardly any LSPs left who are not using your approach is really arrogant and transparently self-serving.”

    This was NOT my post — these are interpretations of comments I made by Dave who actually writes the GTS blog. I tried to correct the tone and intent with my comments under his posting, if you look back at the original post.

    Abbreviation Guide:

    LSP – Language Service Provider or Translation Agency which I agree is more accurate

    CSA – Common Sense Advisory – a company that surveys and gathers market data on the translation industry

    SLV – Single Language Vendor or Translation Agency that specializes in one language

    MLV – Multi-Language Vendor or Translation Agency that project manages many languages for large translation buyers – usually by subcontracting to SLVs or pools of freelancers

    My point was that agencies whose only value add is to provide some basic project management will be marginalized, as other agencies become more efficient either by focus, specialization and more efficient use of technology. The best small agencies distinguish themselves in several ways so I agree with you that they will not be affected by MT or actually are quite likely to use it when it makes sense. I think many that are in the 26,000 entity database at CSA are not likely to survive this shift to efficiency and higher value which often means loyal, high-quality translators who prefer to work with them.

    To be honest my views on the scope of MT have also evolved and today I realize there are places it makes great sense and other places where it would be unwise to use it. Unfortunately there are some enterprise buyers and some MT vendors who feel there is little difference between MT and human translation so this tension and conflict will continue. I think that in many cases MT is used because of volume and urgency rather than cost.

    For a more accurate presentation of my views, http://kv-emptypages.blogspot.com/ is a better representation of my views than the GTS blog you referred to originally.

    Thank you for keeping this discussion open and transparent.

    Like

  16. Thanks for the list of abbreviations.

    1. I put added a link to your blog to my blog roll because many translators are interested in MT.

    2. My suggestion: I would explain most abbreviations the first time I use them. Although these abbreviations are obvious to you, many translators who may happen to read your posts will be lost in them and stop reading if they can’t follow you.

    3. “To be honest my views on the scope of MT have also evolved and today I realize there are places it makes great sense and other places where it would be unwise to use it. Unfortunately there are some enterprise buyers and some MT vendors who feel there is little difference between MT and human translation so this tension and conflict will continue.”

    Amen to that.

    4. Many agencies provide unbelievably sloppy project management. For example, 2 or 3 days ago I received a zipped folders with a bunch of files. I was supposed to choose as many of them as I would be able to translate in the next few days. Another file in Excel had a key as to which files were already translated, but since I misunderstood the e-mail, I picked files that were already translated, after which I was castigate for not understanding what was in the Excel file.

    So I told them that I was no longer available to prevent further misunderstanding between us.

    This is a typical example of “basic management” as it is practiced by many agencies. They can’t even tell what they want from you in an understandable manner, that’s how incompetent they often are.

    Incidentally, since I never worked for this agency (and never will), I did not sign any confidentiality agreement.

    Agencies routinely send dozens of highly confidential files in foreign languages and in English (every page is stamped CONFIDENTIAL) to people who have not signed any confidentiality agreement and who could easily this information for example to the other party in a lawsuit, since it’s quite possible that they are already translating for that other party.

    Like

  17. Thanks for adding me to your blog roll.

    Your experience with this agency is exactly my point,– in this overall group of 26,000 translation agencies that Common Sense Advisory (CSA) has information on, there are probably a more than a few thousand that behave this way. The reality of “project management” for many is no more than to send some (often confidential) information out to lists of translator emails.

    As some agencies become more efficient both in terms of how they use technology and how they interact with translators, this kind of agency should hopefully be pushed out of the market. They don’t deserve to be in business.

    Many of us realize that in the future, there will be a growing awareness that there is a serious shortage of good translators, and finally, the path to being a quality-focused service provider will require collaboration with these good translators, even for those who use MT and other kinds of automation.

    Like


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

Categories

Follow

Get every new post delivered to your Inbox.

Join 1,277 other followers

%d bloggers like this: