FORMAL FALLACY

I’ve seen this argument often lately. And it often goes like that:

^{"If you’re 'good enough', AI won’t replace you."}

or

^{"Translators who don't present (...) skills will be replaced with AI."}

It states that if, as a translator, you’re not specialized, don’t feel linguistic nuances, and don’t put enough effort into making your translation perfect—only then AI is coming for you. But the rest of us can still find our way up.

And I simply can’t agree with that sentiment.

First, the biggest issue is 'selection bias'

What we're seeing is not an evidence of a healthy market. Many skilled and experienced translators are losing clients, rates, or review stages through no fault of their own. Market collapse is real. The pool of professional translators is shrinking.

^{Talented people – really skilled storytellers and translators – leave for other industries. Not because they aren't good enough. Because clients are cutting budgets, choosing automation, or treating localization as a checkbox. The uncertainty about where human localization fits long-term is pushing good people out.}

And the truth is a healthy market would not push the most talented people out. Something bad is happening at the core of the AVT industry and we'll get to it...

Second, the good versus bad translator narrative oversimplifies reality

The idea that AI replaces only lazy or untalented translators is comforting, but false and misleading.

Rates are dropping for excellent professionals. Projects are being removed from human pipelines entirely. Review stages are disappearing. Decisions are made far above linguistic competence. Skill does not protect you from procurement logic and commoditization. Many highly specialized translators are being priced out, not outperformed by better ones.

As localization decisions are driven primarily by price, and treated as a checkbox instead of a strategic choice, this creates an ever filling pool of non-professionals entering and getting out of the trade, not having enough time (or even enough potential—since the price not the talent is the deciding factor) to acquire sufficient skills and knowledge of the sector, therefore further worsening the process of de-professionalisation.

Third, “high end sector” is treated as if it were large and expanding

In reality, that sector is very small and selective. Yes, it exists. No, it cannot absorb everyone who is being pushed out of mid tier and standard work. Calling human expertise a “luxury service” may sound empowering, but luxury markets are narrow by definition. That framing ignores how many careers simply no longer have a viable path forward.

Besides, there's just not enough experienced and skilled tranlators to satisfy the demand for high-quality localization. In fact there's less and less of them. And I'll get back to it later, when we talk about rates.

Fourth, the argument subtly shifts responsibility onto individuals

If you struggle, you are implied to be insufficiently specialized, insufficiently nuanced, or insufficiently excellent. Structural issues like rate dumping, vendor chains, monopolization by LSP giants, and AI first mandates are not being acknowledged enough, if at all.

This allows us to avoid the uncomfortable truth:

And the truth is that even excellent work is often no longer rewarded — it may actually be seen as too expensive.

BEHIND THE SCENES

The biggest shift is not linguistic quality but decision making power. Many clients cannot evaluate TRUE quality and craft, and so they default to price and speed choosing the lesser skilled professionals willing to work more for less money. That is an important structural problem to aknowledge and name.

It’s a common story: companies need a translation or language service, but they often don’t really know what they want. All they care about is the lowest possible price.

Here’s the harsh truth:

Most language-service buyers haven't taken the time to understand the complexity of said language services, from technical accuracy to cultural adaptation and true creativity. They see "translation" as a simple, commodity-like product, not a craft, ignoring the fact that quality varies vastly and cutting corners can lead to costly mistakes and alienation of the audience:

But why is this the case?

Because many buyers focus solely on short-term savings rather than the true value of truly professional, creative and high-quality language services. They treat it like a box to check, not a strategic investment in their brand reputation, global success, and equal accessability (I mean, why shouldn't local players enjoy the same quality as the viewers of the original?).

This approach not only risks delivering subpar content but also damages reputation, audience trust, and enegagement, all for the sake of a few bucks.

The conclusion by some translators that “without us the job just can’t be done” may be comforting, but false

Jobs are already being done without the best talent. Often badly, sometimes barely acceptably (famous "good enough"), but mostly with damage that only appears later. The industry does not stop functioning when quality drops. It just recalibrates expectations downward. Believing that indispensability will automatically be recognized is risky.

Let's be honest. We all know clients and agencies have been cutting corners on translations for years – it didn’t start with machine translations, but when they began outsourcing to the cheapest subcontractors, splitting seasons between multiple translators, accepting machine-like results, and letting typos, grammar issues, and broken translations into the final product. GenAI/MTPE is just the next step in that race to the bottom.

Want examples? Watch the shows or play the games (as much as 85% of them have subpar local versions).

Just look at the screenshots below:

THE WALL OF SHAME

Below is just a small batch examples provided by doctor Paweł Aleksandrowicz, a renowned Audiovisual Translation researcher, in his speech about the declining quality of AVT (you can find the coverage of the Polish version of it here).

Basic grammar errors, typos, clumsy literal translations, even nonexistent Polish words – and nobody at the top cares enough to fix this.

Not just in Movies, in Games too

Here, a unit called 'Skirmisher' was translated into Polish as "Harcmistrz" which means a "Scout Master" in Polish LOL

It should be rendered as "Harcownik" instead, which is a proper translation.

"Raider Leader" and "Foreman" are mistranslated also, though maybe not as spectacularly bad.

^{Game: Drakkar Crew [SiBear Games, 2024]}

Polish version literally says: "Dodge reduces the Stamina by X%". I mean, "great" skill, who’s gonna use it?

On top of that, both Evasive and Dodge have the same Polish translations which just deepens the confusion.

^{Game: Avowed [Obsidian Entertainment, 2025]}

Even the top games have horrible blunders.

Below is an example from the Polish localization of Uncharted. In the original version, one character asks another if they speak English.

But in Polish version, a Bedouin character and Nathan Drake (both with no Polish roots) meet in the middle of the Arabian desert and one asks the other, “If he speaks Polish?” to which the reply is, “Of course I do!”. (well of course!)

I mean, it sounds like the beginning of a bad joke, right? (A Beduin and American walk into a bar…) But not a professional localization. 

Whole games and movies are clearly rushed and/or handed to someone without the skills to save a mere bucks. The system was never about quality, but volume and cost. For years platforms and agencies have tolerated (or actively chosen) bad human translation at bargain rates.

It is not about nitpicking nor achieving an essentially unattainable ideal of every word and every sentence being translated absolutely perfect. The Latin saying: "Errare humanum est...", which translates: "To err is human" is correct, but it also has a second part: "...perseverare autem diabolicum", which means "...but to persist [in error] is diabolical."

Because this is about the general tendency in the localization industry.

Whole games and movies are clearly rushed and/or handed to someone without the skills to save a mere bucks. The system was never about quality, but volume and cost. For years platforms and agencies have tolerated (or actively chosen) bad human translation at bargain rates.

Remember, we aren’t talking about free fan translations here. These localizations are a part of a product that we, as clients, pay for—sometimes hefty amounts—and we have the right to expect the best. 

And while these examples might make us laugh—and they often do—they’re also reminders that behind every clunky line are baffled audiences and creators who lost part of their message.

The sad reality is that quality doesn't follow the money

These figures were provided in a research article by Pablo Romero Fresco, a renowned researcher in the field of Audiovisual Translation. The research was commented by doctor Paweł Aleksandrowicz, on his instagram profile. That's what he had to say:

There is a huge disparity between what a translator brings to the table and what they're able to take away from it. Their profits from the work are disproportionate to the rates they are paid. And this needs to be talked about.

I understand that producers, distributors, and recording studios incur costs, have their obligations, and need to earn a living. But it is worth asking: if a given material is financially successful in a given country, are the profits shared with the translator? Has any translator ever heard: “Thanks to your work, the film or game sold very well, so we are giving you a bonus”? Is the distributor solely responsible for the success of the film, and not the translator?

The truth is that without the translator, there would often be nothing to share from foreign markets money. And if translators are not adequately remunerated, they will leave the profession. And it will not be attractive to young people. Already, graduates with language skills can earn much more in a corporation, while enjoying job stability and not worrying about the hardships of precarious employment. And if there are no professionals and films as well as games are translated by just anyone, viewers will suffer.

And by viewers I mean all of us.

Reality check:

As Jakov Milicevic said:

• Major clients are not demanding lower rates out of necessity; they’re usually influenced by profit motives and the desire to cut costs at any expense.

• Market competition isn’t a force of natural evolution; it’s a tool used by some to maximize profit margins often at the expense of professional linguists and quality.

• The supposed "advances" in AI and automation are being exploited as a smokescreen to justify rate reductions that serve only one purpose: lowering the rates, increasing profit margins for the few at the top.
  
  Instead of investing in true quality, expertise, and long-term relationships, many clients are capitulating to this destructive narrative, building a race to the bottom where vague and superficial profitability is prioritized over professionalism.

SO WILL "GEN-AI" REPLACE TRANSLATORS?

I think it's an irrelevant question. And I can't see that far into the future. But I believe that professional translators will be wiped out sooner by dramatically low rates for our work than genAI.

In fact, it's already started a long time ago. The current financial situation of translators is catastrophic. It is becoming increasingly difficult to make a living from translation alone.

People seem to think that working on big American productions, with budgets in tens and often hundreds of millions, means great earnings. And maybe it used to be that way. Unfortunately, for 20 years we have been witnessing the industry's steadily deepening decline, widespread contempt for our profession, and constant pay cuts.

As a result, most translators are unable to earn even the minimum wage. Just look at the research on audiovisual translators' working coditions:

The reality is not encouraging. GenAI only exacerbates the gloomy mood. Source: AVTE

"If it's so bad, why won't translators quit?"

I think it's a wrong question. Because a healthy industry wouldn't push the most skilled people out.

But also, the truth is, people go into game localization and audiovisual translation for all the wrong reason. Mainly, out of passion for games and movies, and that passion can be easily exploited, which leads to lower pay, longer work hours and eventually a burnout. And this is a case with many experienced translators, many of whom have already left the traslation industry for greener pastures. I started my career over 22 years ago and ALL of my peers from that time are no longer in the business of translating.

So you shouldn't go into Game Loc and AVT because you like Call of Duty and Stranger Things. You should pursue this career if you have talent for it and you're good at it. And then, if you know your worth, you should value yourself accordingly, not accept low rates and participate in price-dumping.

Because now you may be a cheap beginner, but it's gonna be extremely hard for you in the future, when you have bigger responsibilities and bigger expenses (maybe even a child), and you won't be able to raise your rates even above the rate of inflation. Don't believe my own experience? Just look at the research then:

AVTE research shows that experience has no discernable impact on earnings. Okay, if you're a newbie translator, you may not earn much; that's just how the market works. Maybe you're not very productive, maybe your work still needs a lot of editing. That kinda makes sense.

However, it is not better when you have between 6 and 19 years of experience, nor is it much better when you have over 20 years of seniority. Still more than half of these professionals earn salaries that puts them in the lower half of the economy.

For me, this is horryfying. Imagine working for 20 years in one profession and still earning virtually the same as a beginner. And I'm not even talking about the amount itself, because we know that there is inflation and these rates need to cover that inflation over time—so your purchasing power and financial status remain the same. So, even though you are no longer a newbie, but a highly qualified professional with experience who has seen a lot in the industry, you're still gonna earn the same as at the beginning.

And so this line of work doesn't attract the best individuals anymore, but the cheapest and often the amateurs willing to work for peanuts.

Truth is, as in all other industries, the only way to attract top talent is not with the superficial promises of "satisfying work, Pizza Parties and great vibes", but to pay those professionals accordingly to their skill level and experience.

And you can argue as much as you want, there's no way around it.

So are translators afraid of AI?

Those who have honed their craft over many years know that algorithms and language models cannot cope with translations into Polish. And certainly not with narrative game localization and audiovisual translations, which require natural spoken language and in dubbing synchronization with the characters' lip movements. The material we deal with on a daily basis requires such enormous flexibility and experience that, from our perspective, machines do not pose a threat.

The problem is that it's not translators who get to say who will work on a given project and how, even when we already did previous installments of a movie or a series. This is usually decided by clients and agencies, whose main goal is often to get the translation done as cheaply as possible. This is a strange strategy, considering that film production itself is extremely expensive and saving money on it at the last stage is a classic case of shooting yourself in the foot.

The only hope lies with viewers who, upon noticing the poor quality of the translation, may not only stop watching the work in question, but also express their opinion on the internet, which I strongly encourage to do.

^{Films and games are made for audiences, and—contrary to appearances—their voice is very important to producers and developers. If audiences accept machine translations, won't see the difference, and won't say no to AI-slop, then translations with a human element will simply disappear.}

THE END — LITERALLY AND FIGURATIVELY

What is missing entirely is collective action or systemic response.

There is no discussion of rate floors, client education beyond individual consulting, industry standards, or refusal strategies that actually cost decision makers something.

The solution for movie producers and game developers is to move localization upstream. Treat it as part of the production, not post-production. Have translators work directly for you. They you still have them work through the agency and tools of your choice but they should be treated as client-translators who work for YOU, not as agency-translators who can be exploited and burned-out behind the curtains.

Unfortunately, that's exactly where there's resistance. Clients won't spend any money on localisation because executives and accounting see it as a cost rather than marketing or investment.

The biggest problem for localisation is not that AI exists. It's that localisation is funded from the wrong budget — as part of post-production, and not a production.

Meanwhile, good localisation would cost way less than one regional launch party or social media campaign, or even advert...but guess which one gets pushed on the price?

The result is, we all suffer, both skilled translators who are pushed out of the market, and viewers who get the increasingly more distorted version of the original work.