Tag: AI

More AI Snake Oil

In a recent BBC report of an interview with Google (Alphabet) boss Sundar Pichai, we see Pichai repeating the nonsense spewed by numerous AI tech bros in attempts to allay fears of AI disruption.

Pichai says that AI will also affect work as we know it, calling it “the most profound technology” humankind had worked on. Well, perhaps that part of his comments is correct.

“We will have to work through societal disruptions,” he said, adding that it would also “create new opportunities”.

One person’s manageable disruption is another’s total disastrous loss of earning power.

There are people who have been successful in a number of specific careers who will not be offered or be able to take those “new opportunities.” Many are already being deprived of work and income, by being replaced by AI. Opportunities will be provided to others, perhaps, but certain careers will simply disappear; translation is one of them. Change is coming much too fast for some groups to keep up. The notion that translation will continue as a career is clearly delusional.

“It will evolve and transition certain jobs, and people will need to adapt,” he said. Those who do adapt to AI “will do better”.

Transition? Adapt? What do translators transition to and how do they adapt? By accepting post-editing work at one-fifth the word rate (and certainly not anywhere near a compensating five-fold increase in throughput)? And that doesn’t even address the issue of mind-numbing post-editing work. Most translators will not be able to survive by adapting to AI or even adopting AI themselves, because they have been habitualized by their clients to working only for a client demographic one tier above them on the food chain, and that client demographic is currently using AI to replace them. Freelancers have arguably allowed themselves to be isolated from translation consumers that have not yet adopted AI. AI will be of no avail without clients that are not using AI, and that includes acting as the “human in the loop,” the snake oil sold by the major translation brokers that control most of the translation market.

“It doesn’t matter whether you want to be a teacher [or] a doctor. All those professions will be around, but the people who will do well in each of those professions are people who learn how to use these tools.”

This could very well be true for some professions. As noted above, however, freelance translating is not one of them, as is already being demonstrated to be the case, with large numbers of translators trapped into non-translation functions that pay extremely poorly, if they even have that left as their clients replace them with AI.

The outlook is bleak. More and more translators are coming to realize that, but numerous translation organizations appear to hang onto the delusion that this storm can be ridden out; it cannot. When it passes, it will leave a barren wasteland in its path, with very few translators left standing.

Falacious arguments are used to support delusions of human translator survival.

Even as AI translation continues to make progress in replacing countless Japanese-to-English translators, two falacious arguments continue to be used to support the mistaken prediction that AI won’t replace professional human translators.

  • Translation by AI cannot capture cultural nuance and emotion.

This is arguably true, but for only a tiny portion of translation that is paid for. Only a very small portion of the translation that is paid for has anything to do with culture or emotion. In contrast to the relatively much smaller fields such as literature, games, other entertainment, or marketing, translation in fields such as finance, business, technical, legal, patent, and the like, which have supported the livelihoods of the overwhelming majority of Japanese-to-English translators do not involve culture. There’s just not much cultural nuance or emotion in sight in most of the mainstream translation fields, which can be handled satisfactorily—the client decides what is satisfactory—by AI combined with low-paid post editing.

  • Mistranslation will result in injuries, deaths, and litigation.

However attractive such outcomes might seem to translators hoping for a schadenfreude hit, these warnings, almost always presented in the abstract without factual basis, are not supported by verifiable incidents. These warnings of incredible risks need credible facts to support them. Those facts haven’t been forthcoming.

No amount of chanting of the cultural-nuance mantra or the deaths-and-litigation mantra will change what has already progressed in translation to an extent not much talked about, perhaps because the people affected are gone.

The replacement of professional human translators by AI with post-editing is well underway and coming soon to your optimistic neighborhood.

A Non-Crisis

There has been a kerfuffle recently in Japan surrounding a term that can be translated as existential crisis. It revolved around the newly elected PM’s mentioning that an attack by China on Taiwan could be taken to be an existential crisis, apparently for Japan, although that was not explicitly stated. This led me to think about something closer to the lives of translators that could be called an existential non-crisis.

Some translators might think that the appearance of AI has created an existential crisis for human translators. I disagree, because the use of the term crisis implies a possibility of surviving the crisis. For almost all freelance translators—and arguably many translation companies, although a bit later—survival as a translation provider isn’t in the cards.

It is time for both translators and their organizations to stop pretending—either by studied silence or by active denial and diversionary handwaving—that human translators are not already being replaced by AI on a devastating scale or won’t shortly be almost totally eliminated from the translation process step.

It is also time to realize that the essential and immutable enabling condition for the demise of human translators by AI use is the two-tier brokered structure of freelance translating, and that this condition was satisfied decades before AI appeared. That structure welcomed and fostered a population of freelancers who are willing to be isolated from translation consumers—some even welcomed that isolation—and who are incapable of surviving when the only client demographic available to most of them decides to replace them with AI.

Continuing the pretense of resilience in the face of the AI transformation might feel good (or at least might feel less painful than the truth), but it won’t change things.

Let’s get real and look at the real future, which will be radically different from what most translators imagined when they started translating. For most current freelance translators, that future won’t involve translating.

Are translators’ organizations giving this any thought? It’s hard to tell. One organization I am familiar with is basically silent on this situation. They hold occasional events on topics of interest to a small group of translators in a limited subject-matter domain, but they appear to have lost relevance to many members, judging from their significant loss of members, starting in 2016, long before the pandemic, which has sometimes been cited as the cause of the drop in membership.

I placed this post here because I suspected that, placed somewhere else, it would evoke serious pushback, denial, and even anger from the very people who will shortly be—or already are—ex-translators. Pushback, denial, and anger won’t change things.