2023-05-19 06:50:08
“My love grows stronger, if its appearance is weak; / and although it shows it less, I have not stopped loving you; / Merchandise is the public love whose excellence / the tongue of its owner proclaims anywhere.”, says one .
“My love is strengthened even if it seems weaker;/ I do not love less even if I show my love less;/ That love is merchandise, whose rich appreciation,/ the owner publishes everywhere with his tongue.”, maintains the second.
And, the third, affirms that “Although weak in appearance, my love is very strong,/ and I do not love less, just because it seems so./ It is mercantile love, the one whose high esteem,/ everywhere is proclaimed by the language of its author.”
It would seem easy to distinguish which of these versions are the work of a human being and which of a artificial intelligenceAfter all, it is about poetry, the most difficult to translate. At a time when newspaper articles abound reflecting both the fascination and fear of artificial intelligence, the translationsthat minefield of the labor editorialseems perhaps the most propitious environment for its use and for, incidentally, Reduce costs and get rid of the everlasting translators’ complaints.
For Jorge Carrionco-author with Taller Estampa and GPT of Electromagnetic fields. Theories and practices of artificial writingtranslation programs are “very useful tools for obtaining a first version of scripts, technical or popular books, as well as essays without literary ambition. But then you need a human supervision. they still can’t translate high literaturebut it is likely that soon they will be able to do so.”
Of the same opinion seems to be pia gepetto (name we have given to the ChatGPT for your participation in this report): “Machine translation programs have come a long way in recent years, and in many cases can produce literary translations that are accurate and understandable. However, the Literary translation it’s a highly specialized field and often those programs are not able to capture the deeper sense of languageincluding nuances, puns, double meanings and metaphors, key elements of literature.”
For Ernest Folchpublisher of Navona and Folch&Folch, it is only “vaguely possible to translate highly technical and highly mechanical texts with AI programs, something like a prescription or a traffic ticketAnd yet I’m skeptical. A translation, like any creative work, needs things like emotion, sensitivity and good judgment, all virtues that a machine can never have.”
Ernest Folch, editor: “A translation, like any creative work, needs things like emotion, sensitivity and good judgment, all virtues that a machine can never have”
Magdalena Palmerwinner of the XXIV Ángel Crespo Translation Award, also thinks that AI translation “is a useful tool for technical translations, but its application to literature is a mistake. And always, even if they are technical translations, they have to go through revision and the supervision of highly experienced human eyes to be able to detect and correct serious errors.As for literary texts, there are so many factors that influence the decision of how to translate a phrase, an expression, a metaphor, a game of words, the register of dialogues, a character’s way of speaking, cultural biases, etc., which I don’t see as possible without seriously diminishing the quality of the text”.
reduce costs
Even so, he sees it as possible that “some publishers try to get us translators to become proofreaders of translations made through AI. Paying us less, of course. The application of AI to literary translation will aim to reduce costs, not improve the quality of the translation”.
Magdalena Palmer, translator: “AI is a useful tool for technical translations, but its application to literature is a mistake”
For the translator Cristina Maciatranslating literature by machine is “unfeasible” and “all attempts have ended in catastrophe”, although, if it were ever possible, imagine a future in which there are four highly valued translators left in charge of deluxe editions sold with slogans such as “100% manual translation!”, almost as if it were an ecological product without additives.
victims
In any case, although most of the reactions of other consulted publishers oscillate between incredulity and horror, in the gossip of the industry there is a whispered mention of a publisher that translates its less demanding books through software and then corrects them and launches them on the market. If it were true, it would be a bit like that saying about witches: books translated by AI don’t exist, but there are, there are. And its first victims would be the translators, but also the readers.
In the gossip of the industry, a publisher is mentioned in a low voice that translates its less demanding books using software
for the writer Patrick Pron“not all publishers are sensitive to the fact that a good translator is a commercial claim for your books, in addition to a quality guarantee; those publishers pay little or very little to their translators, and the fact that from now on they have virtually no pay is great news for them, but not for those who will have to compete with those publishers. digital slaves in semi-slavery conditions nor, of course, for readers, who will be paying for worthless books”.
Distinction
For Carrión, the moment in which it will be impossible to distinguish whether the translation was done with neurons or with bits will undoubtedly arrive, “if it hasn’t already arrived and we haven’t realized it”. The nice Pia Gepeto says that “it is possible that in the future machine translation technology could reach a level where it is difficult to distinguish between a translation done by an AI and one done by a human being. However, this will largely depend on the quality and complexity of the literary work in question, as well as the level of cultural and historical knowledge that is required to fully understand the work.” At that point, Pron says, “I no longer it will matter a lot who translated a book or who wrote it; for that reason, it will not be necessary to read it either”.
Perhaps the great danger hidden in automatic translations is that, at some point, they will become the standard of quality
In a recent article in the New York Times titled “The False Promise of ChatGPT”, Noam Chomsky maintains that the great fallacy of AI It’s because “the human mind is not a heavy statistical pattern-matching engine, devouring hundreds of terabytes of data and extrapolating the most logical retort in a conversation or the most likely answer to a scientific question,” but “a surprisingly efficient and elegant system that works with small amounts of information and that does not seek to infer crude correlations between data but to generate explanations” (translation is mine).
Neoliberal strategy
Publisher Ernest Folch would probably agree. For him, we are facing a fanaticism of technology and a neoliberal strategy to save costs. “They want us to believe this stupidity that machines can also replace and even improve our brain. And the truth is that a translation, like any creative work, needs things like emotion, sensitivity and good judgment, all virtues that a machine can never have. There is nothing further from literature than a machine, and I honestly don’t believe that there can be literary translations, much less decent and intelligent literary works that can be produced by AI. So far the only big find of the AI is the name, which gives it an appearance of what it is not. It may be useful for many things, but definitely not for anything to do with creation.”
In any case, perhaps the great danger hidden in automatic translations is that, at some point, they will become the standard of quality, and that, in the not too distant future, publishers require their human translators to translate like machines. There there will be no Terminator to save us.
[Nota: la primera de las traducciones del soneto de Shakespeare del principio de este reportaje pertenece a William Ospina; la segunda es de ChatGPT y la tercera de Ramón García González].
#Translation #check #artificial #intelligence