Whatever the Wind Brings

Using "AI"-translated content is a bad idea

In some ways, I could be considered a Luddite, I guess

I'm not against it because I fear for my job. I actually would embrace any new technology that made things easier, but that hasn't been the case here. The issue with "AI" – and I use quotes because it is not a true AI, despite its hype – is that it is not intelligent: it doesn't understand context or meaning, it just makes decisions based on statistical modeling. That's not how we work with words. And that's as bad as it is unpredictable.

This tech not only has been devaluing my profession because tech companies are saying it works perfectly and doesn't need more human translators, but it also makes my job much more difficult while paying less for more work. The math doesn't add up on my side. If the technology isn't making me more "productive" and outputting the same quality, why should I be charging less for what I do?

In any case, the company I currently work for made the decision, a few years ago, to stop accepting MTPE jobs – "post-editing" a translation made by a machine. I could give tons of reasons why this was a smart move, but it's tiring to keep repeating the same things over and over again. Luckily, there's a recent article by IGDA that summarizes perfectly why it is a bad idea, and why "AI" isn't a solution for anything. If you're interested in the topic, I recommend taking a few minutes to read it.

Check the article clicking here.

The funny thing is that almost every translator I know, who works in this field, arrived at the same issues and same conclusions, and has been warning people for years about it. This wasn't an "organized movement" until recently – and to be honest, it still isn't, it's more like a PSA from experienced folk in the field. So when I'm asked if I'm afraid of losing my job to an "AI", I always answer "No, because it still can't do what I do. The real issue is convincing the executives about that".

And funny enough (again), I've seen many translators, who don't know each other, giving the same answer.

#random