Why choose a human translator?

A friendly humanoid robot looks at the camera

WHY CHOOSE A HUMAN TRANSLATOR?

In the age of DeepL and ChatGPT, why should you choose

me to translate or proofread your text when you could use

AI for free?

1) Why would people want to read something that nobody could be bothered to write?

Translation isn't as simple as just looking up words in a dictionary. Very often, it's about crafting a new text for a new target audience. And why would people take time to read a text that nobody made the time to write properly? It's simple: if you wouldn't use AI to write the original text, don't use AI for the translation.

2) Machines can't translate context.

Just like in English, most words in other languages have more than one meaning, depending on the context. So, to make sure your translation makes sense, choose a human translator. For examples of epic translation fails, click here.

3) AI isn't good at humour or word play.

Got a marketing text full of puns and jokes? Don't get it translated by AI. At best, it won't be be funny. At worst, it could end up being downright offensive...

4) Machines can't explain cultural differences.

Every language comes with at least one different culture, legal system, education system, social welfare system (or lack thereof), and many other aspects. To do their job well, translators sometimes need to add information to explain the differences in culture or system to your new target audience. This is something a robot won't do for you.

5) Machines can't cope if there's a mistake in the source text.

Most texts contain at least a few errors (a missing word here, a grammar error there, a word the spell check has auto-corrected by mistake...). As humans, we can usually understand what's meant anyway. But machines can really struggle with this, depending on the kind of mistake and the language involved - sometimes with disastrous results! If ever I am unsure of what your text means, I will ask you to make sure! I will also point out errors in your source text.

6) Machines don't adapt to changes in target audience.

Every translation involves a change in target audience - be it from a German-speaking audience to a British English-speaking one, or from an Arabic-speaking Syrian audience to a US English-speaking one. This means that words such as "foreign" or "abroad" need to be repositioned to reflect the location of the new audience. Again, this is something only a human translator will do.

7) AI might be quick and free, but often even the best it can produce is pretty mediocre.

In the words of Mark Twain, "the difference between the almost right word and the right word is really a large matter -- it's the difference between the lightning bug and the lightning."

If you need to use precisely the right word every time, contact me for a quote!

8) AI feedback loop/model collapse
The more people who use AI, the more AI will be fed with and trained on AI-generated data. This could lead to a so-called feedback loop, or eventual model collapse, as the quality of material gets progressively worse due to a lack of original, human input. If you value originality and the human touch, hire a human translator!

9) Data protection
Any data you enter into Google Translate is theirs for the taking.

10) If you don't want to take my word for it, this is what ChatGPT said when I asked it about its limitations:

"While ChatGPT can be a useful tool for various language-related tasks, including translations, it has some limitations. Here are the key limitations you should be aware of when using ChatGPT for translations:

Contextual Understanding: ChatGPT processes text in a context window, which is limited in size. It may not fully understand the context of a long conversation or complex sentences, leading to potential mistranslations.
Ambiguity Handling: ChatGPT might struggle with ambiguous phrases or words that have multiple meanings. It may provide translations that are contextually incorrect or incomplete.
Accuracy: While ChatGPT can generate human-like responses, it doesn't have perfect accuracy. Translations might not be as precise as those from dedicated translation models.
Domain-Specific Language: If your text involves specialized terminology or domain-specific language, ChatGPT may not accurately translate such content, as it may lack specific knowledge in certain fields.
Cultural Nuances: Translating involves understanding cultural nuances, idioms, and expressions. ChatGPT may not always capture these subtleties accurately, leading to potential misunderstandings.
Inability to Learn From Corrections: Unlike a learning model, ChatGPT doesn't learn from corrections. If a translation is incorrect, providing feedback won't help improve future translations.
Lack of Real-Time Updates: ChatGPT's knowledge is static and does not get updated in real-time. Therefore, it may not be aware of recent events, changes in language usage, or new terminology.
Sentence Length Limitations: There are limitations on the length of input and output sequences that the model can handle. Very long sentences may get truncated, affecting the quality of the translation.
Source Language Proficiency: The quality of the translation might depend on the proficiency and clarity of the source language input. Unclear or poorly structured sentences may lead to less accurate translations.
Unsupported Languages: ChatGPT may not support all languages, and the quality of translations might vary across different language pairs.
Data Bias: If the training data used to train ChatGPT contains biases or imbalances, these biases may be reflected in the translations it produces.

It's important to consider these limitations and use ChatGPT for translations with a critical eye. For critical or professional translation tasks, dedicated translation models or services may be more suitable."

Image: Alex Knight, Unsplash

Contact me