Long briefs and AI wisdom

 

A word to the wise is sufficient, but only sometimes; a kick in the ass makes a far greater impression. As a case in point, this week I made an apparently unsuccessful proposal on a translation project. In retrospect, I could have and should have handled it differently and will do so in the future.

The project in question was the translation from French into English of a 45-page real estate purchase contract containing nearly 25,000 words. The document posed several challenges, including PDF formatting and quite precise legal vocabulary. Furthermore, the customer required a signed statement of accuracy, which would entail even more QA time to ensure that demand. Due to the extensive time required because of the document’s length and complexity, the quote involved four figures (in euros) and a long deadline. I received a polite but understandable “I will let you know when relevant”. It was clear from the start that the customer was neither expecting or willing to pay such an amount. To be fair, I was not surprised but stood by my quote as I knew how much time and effort such a translation involves.

Looking back, I see that I had another option, one that I know about and had forgotten. Specifically, I had heard Naomi Shenkar, a respected Israeli attorney and translator, say that, for long documents,  she provides two quotes: one for human translation and one for AI + editing. The first is based on standard human translation elements while the second quote is contingent on the quality of the AI output. She also noted that most customers ultimately prefer the human translation. In this matter, she provides the potential buyer with a choice, an unequal one in many respects, but at least an option.

I now recognize that I should have done the same. If I had offered to edit an AI version, the customer may have considered and understood the human translation proposal. While I am not particularly fond of AI or any machine translation, for longer documents for which the time and financial investment may be beyond the customer’s situation, it can provide a reasonable alternative. Furthermore, I would have created a win/win solution. Shorter projects create income and allow the translator to help other customers.

When I began translating some 22 years ago, I promised myself that I would limit myself to one time per mistake. I have been fairly loyal to that vow. The next time I have the opportunity to make a proposal on a long legal document, I will provide an AI option. Better late than never.

Prime perspectives on adopting a new country

 

37 years ago, on the date I am writing this post, I made aliyah, immigrated, to Israel. The actual day was not as eventful as D-Day, which also occurred on the same date, albeit some 54 years previously, but has had a huge impact on my life. I came to Israel and have made it my home, notwithstanding 5+ wars, because my decision was carefully calculated, both rationally and emotionally. I have paid a price in terms of becoming an alien to one degree or another. However, looking back, it was the best road taken.

Every immigrant’s story is personal and unique. In my case, moving to Israel was the search for a place that felt permanent. I had spent most of my adult life moving from city to city, apartment to apartment and job to job. I enjoyed Oregon and still appreciate its pace and beauty. However, it never felt like home despite the friends and lifestyle. Facing another period of unemployment, I chose to join the yearlong “Sherut La’am” [Service to the People] program in which I taught English at a high school in Ashkelon, in southern Israel. We lived in and taught students from the “wrong side of the tracks”. Many of my neighbors had no books in their houses. Even worse, one neighbor was a drug dealer while there was a local “working lady” below in the square. We received housing and a small stipend. I was contented for the first time despite the less-than-glamorous reality around me. I returned to the States for a year and then formally decided to immigrate to Israel. My father asked me: “Are you going to or running from”. The answer was clear and correct. I was going to. I ended up in Karmiel, in the North, where I still live today. I have never looked back.

My successful transition to living in Israel, far from obvious, derives from my decision to embrace the whole experience, positive and negative, to the best of my ability. I insisted on speaking Hebrew and actually avoided the Anglo-Saxon community. I married an Israeli and raised my child in Hebrew, watching all her videos and reading her books in Hebrew, mistakes and all. My second wife is also Israeli. I have tried to understand and selectively adopt Israeli social mores or at least anticipate them. As they used to say, I went native. I did not become fully Israeli nor will I ever be. I arrived too late in my life and did not serve in the army. Furthermore, it is impossible to erase almost 30 years of education, formal and informal. Yet, I am much as part of Israel as I can be.

Of course, there was a high price to pay: becoming a permanent alien. I do not speak nor in particular write Hebrew as a native language. I miss small subtleties in conversation and make stupid errors, particularly with gender forms in verbs. Culturally, I have learned to expect certain behaviors but cannot or choose not to adopt them. They are not my style. Not having grown up here, I don’t have friends from childhood or the army and lack the “bande”, as my mother would say, the group of people that have known you since you were small. For the same reason, I lack the shared experience of high school, the army or university studies. My growing-up period was in a different country. On a deeper level, I also stopped being an American. The America I grew up in has long disappeared. I felt the divide even while I was still frequently traveling to the States to see my parents. I can live in three countries (the United States, Israel and France) without belonging 100% to any of them. That loss of clear identity is a heavy price.

Yet, as Edith Piaf said, Je ne regrette rien – I regret nothing. Even before the current ugly wave of antisemitism in the United States and Europe, Israel was my home. It is an imperfect place, I would agree, but native born Israelis are equally, if not more, critical than I am. I am always happy to return to my country, wars, missiles, and political chaos included. Ultimately, without fully being aware of it, Israel is my home because I am Jewish here without apology and adjustment. I do not have to prove my Jewishness by being religious or politically active nor do I have to push back against my neighbors that are actively or passively antisemitic. To highjack Jacques Prévert, je suis comme je me suis; je me suis fait comme ça, or, in English, I am what I am; I made myself that. I have become Israeli in my own way, just as hundreds of thousands of immigrants have done before me and after me. Even more importantly, home is where you feel internal peace. I have always felt at home in Israel, from my first day here. I do not wish to live anywhere else. Thus, on this 37th, a prime number by chance, anniversary of my Aliyah, I have attained my goal and am content. That feeling is the ultimate success.

Google-eyed translator search – Shocked and uh-ed translation buyers

 

Freelancer marketing has never been a simple task just as finding any proper professional for the first time has always been a bit of a gamble. As an exercise, I “posed” as a translation buyer in my area of specialty – Hebrew-to-English official documents, a marriage certificate in this case – and searched for a service provider using both ChatGPT and Google. While the former provided a reasonable response, the options raised by Google illustrated the irrelevance of the search for translation purchasers and the need to focus marketing for translation providers.

In terms of methodology, I used free ChatGPT and posted the following prompt: “I am looking for a certified Hebrew-English translator to translate a marriage certificate for immigration purposes.” I then added, “I am looking for three freelancers”. Aside from some fairly accurate but somewhat vague background information, which is understandable since my prompt did not include all the relevant information, the AI engine initially provided three agencies and then, following the second prompt, three freelancers. The referred translation agencies and independent translators were all experienced professionals. Thus, GPT did provide an adequate starting point for a search for a translation professional.

By contrast, Google’s output to the same initial question was highly biased. The first page was almost entirely AI-based translation companies. I understand that people pay for top billing but I am not familiar with any government agency that will accept a certified translation of an official document produced by AI. The first human agencies were a few large, low-cost agencies. I am aware that some customers have a limited budget but it seems penny- wise and pound-foolish to skimp on the translation of a document that may make or break an important opportunity. Among the early entries (page 2) was a site for freelancers. It was shocking to see their average prices. Almost all of the more than 20 freelancers on the first two pages posted a rate of 5 USD a page, without defining what a page is. At that price, I would not trust the result as the rate is too good to be true. Thus, faced with a choice of AI, budget agencies and prostituting freelancers, the search was a non-starter.

It would seem that first-time translation buyers seeking a professional translator should take a much more directed approach. The best option may be national association search engines (such as the ATA in the US and ITA in Israel), which allow a full comparison of credentials and demonstrate some measure of professional approach. Service provider recommendation lists may also be relevant, albeit with a grain or two of salt. Of course, traditional personal recommendations are always relevant. Notices in relevant Facebook groups will provide several options. Finally, ChatGPT with more specifics is an ever more relevant path. Clearly, finding an appropriate translation will require some thinking and effort.

This situation should determine how translators market themselves. As AI seeks specifics that make one option more relevant than the others, translators need to include relevant details in as many locations as possible. These factors include languages, qualifications, experience, types of documents and skills. A website is an important but not the sole place to post such information. Furthermore, appearing on reputable translator lists, e.g., national associations and government bodies, enhances the probability that a search will produce one’s name. Of course, additional activities, including lectures, blogs and volunteer work, further enhance one’s presence. Ultimately, word-of-mouth, whether through customers or peers, is the best source of reference, requiring translators to produce quality results regardless of the rate. It is not enough for translators to prove and improve themselves. They need to market those attributes in multiple ways.

God helps those who help themselves. I am not sure that Google is a god even if it does seem to be omniscient and omnipresent. AI is far too naive to be a god but people increasingly treat it as one. Translation providers need to help purchasers find an appropriate option. Keeping on the astral plane, the devil is in the details. Purchasers need to make intelligent decisions while freelancers must provide the information they need to make those decisions. The result can be awe-inspiring, far superior to the uhs produced by simple Google searches and even ChatGPT.