AI in translation. My two cents
Yes, this is another AI post.
Like you, I’m also a bit overwhelmed (and tired) by all the hype about AI tools.
But we cannot entirely escape from it, so I try to get as much information as possible.
At my own pace.
And with a pragmatic approach.
Of all the courses, webinars and offers available, I decided to attend the AI in Translation Summit.
First, because it was affordable.
Second, because the talks were short, and as I’ve purchased the Power Pack, I can listen to them in between projects (instead of watching pet videos on Instagram).
I haven’t seen all the videos yet, but this post is my view of the current status of AI in translation and how it can affect my business.
Is AI really that new to us???
The more I read about AI, the more I thought it was nothing new.
Machine translation has been around us for more than 30 years already!
The current model, neural machine translation (NMT), is a form of AI.
So, in terms of being scared by new technology, I’m not.
The translation job market in the AI era
This was one of the main reasons why I joined the summit.
I somewhat agree with Daniel Sebesta.
Some translators will be out of business. Those who don’t accept the trends.
There are two trends:
- Translators becoming post-editors or pure linguists.
Haven’t you noticed how translation requests have been reduced in the past 10 years due to machine translation?
In some domains, translators will work more and more as MT post-editors or pure linguists, helping develop new AI translation products.
There will be other more critical fields where human translation will still be needed.
- Translators becoming language consultants.
We have always been language and culture consultants.
With more technology options available to our clients, our work is also to advise on how to make the best use of technology and help them choose the content they have to translate and the best approach (human, NMT, hybrid).
According to Daniel, those who don’t accept these new scenarios and who don’t work in domains that guarantee good volumes and good rates to make a living out of it will leave the business.
I’m not that sure, but it could be possible.
Using AI in translation projects
Arle Lommel made a very interesting comparison between the current status of NMT and AI.
Currently, NMT is better than AI when it comes to translate texts. As I said before, NMT is an AI, but with some capabilities that AI still lacks.
In my opinion, AI is good mostly for long-form content translation, like blog posts, guides or training materials.
And after translation, you can use AI to polish it. For example, to find synonyms or check for long sentences and shorten them, as Rachel Pierce showed in her talk.
However, I don’t see this as feasible for short translations and high turnaround times.
Will AI make us lazy translators?
I’m also reading this a lot on LinkedIn, particularly among content creators.
I think AI is good for getting ideas and repurposing content. For example, I’ve used ChatGPT to shorten a LinkedIn post that was too long.
But I didn’t take all the AI output. Just what was suitable for my post.
Also, all that hype about AI being the best that could happen to us is nonsense to me.
Are you going to trust everything you do to AI?
Where are your translation skills?
MT makes me a worse translator.
For example, when I work on MT projects for a long time and switch to a marketing translation, my translation is not as natural as it should be.
I need some time to warm up and get the creative juices flowing.
That’s why I limit the amount of MT projects I accept.
I don’t trust everything I do to AI. I’ve been working in translation since I graduated in 2001, and I’m not going to put all my experience and constant training aside just for a fancy tool.
AI is another tool available to us. It’s not THE tool, and it’s not a replacement, but an aid.
We can’t go against technology.
And trying to keep up with everything is overwhelming. At least for me.
While I agree with much of what I’ve seen in the summit, some clients prefer to work with human translation. I know it because I work with them.
The only ‘problem’ is that these clients aren’t easy to find.
But, hey! Who said finding clients requiring translation and willing to pay fair rates was easy?
2023 has been the year when I’ve decided to work with quality clients.
I have been saying this since I started freelancing in the summer of 2016, but, like you, I was always afraid of saying No.
But this has proved to be my best strategy.
Since I said No to clients who didn’t appreciate the value of my work, I have better clients.
I’ll continue with this approach, and diversifying my services.
I’ve always been a consultant to my clients, trying to keep up with the latest trends and developments to advise them on the best translation approach.
I’ll continue doing it, now exploring the possibilities of AI for them.
As I do with any other technology.
I hope this analysis helps you better understand the AI landscape in translation.
Maybe I’m too simplistic, but I like simple things.
See you for the next post!