Leonid Glazychev,
Logrus IT’s CEO
I still remember a terrifying moment from my university days. I was in a professor’s office – a small room dominated by a huge, neglected aquarium filled with disgusting, murky brown water. While we were talking, the professor, without even looking, tossed a cigarette butt over his head. It flew in a perfect arc and landed right in the tank.
Immediately, the murky water started to boil. Unseen "monsters" surfaced and devoured the cigarette butt in seconds. It was like a Hitchcock movie; the fact that you couldn't see exactly what was happening below the surface made it even scarier.
Why am I telling you this story? Because for the last 30 years, "Pure Human Translation" has been the gold standard – reliable and trusted. But today, the commercial reality in our industry started to resemble that murky aquarium. We are living in an illusion!
The Economic Reality. Translators are highly educated professionals required to know multiple languages and complex subject matters. Yet, for decades, rates have been stagnant or decreasing.
When you combine high expectations with low pay, the temptation to cut corners becomes extreme. Whether it’s skipping the review phase, outsourcing to unqualified amateurs or non-native speakers, or using unauthorized Machine Translation (MT) or AI – deviations are happening constantly.
The "invisible" AI. You might ask, "Why not just check for AI usage?"
Of course, we do it at Logrus IT! We run spot checks and QA. But here is the hard truth: AI is getting too good to catch. Modern AI tools can often improve raw MT and/or mimic human style so well that distinguishing between a "pure" human translation and an AI-assisted one is becoming impossible. It is no longer verifiable, only the most lazy or stupid violators are caught in the act...
It’s like marathon runners are taking a train to the finish line in big numbers... They arrived at the destination, but they broke the rules of the competition.
So, if "Pure Human Translation" is now unverifiable (and hence not viable) on a commercial scale, what is the alternative? I’ll be discussing the solution in my next video. Stay tuned.
