| Pages in topic: < [1 2 3 4 5 6 7 8 9] > | Consider the consequences of AI before you start a career in translation Thread poster: Gerard de Noord
|
|---|
| Artificial... | Apr 2, 2022 |
semi-stupidity?
German-Dutch Engineering Translation wrote:
I know that almost everyone uses the term 'artificial intelligence' for DeepL and the like. But I don't find much 'intelligence' in it. It is about pattern recognition and statistics.
| | | | Daniel Frisano 
Italy
Local time: 10:13
Member (2008)
English to Italian
+ ...
German-Dutch Engineering Translation wrote:
I know that almost everyone uses the term 'artificial intelligence' for DeepL and the like. But I don't find much 'intelligence' in it.
Spot on. They have to sell it somehow, don't they?
You don't sell sugared and caffeinated water by calling it sugared and caffeinated water. It needs to be called an "energy drink" or something.
| | | | Dan Lucas
United Kingdom
Local time: 09:13
Member (2014)
Japanese to English
Giovanni Guarnieri MITI, MIL wrote: semi-stupidity? German-Dutch Engineering Translation wrote:
I know that almost everyone uses the term 'artificial intelligence' for DeepL and the like. But I don't find much 'intelligence' in it. It is about pattern recognition and statistics.
Author Linda Nagata used the phrase "dull intelligence" ("DI") back in the 1990s when she wrote her far-future novel "The Bohr Maker". I don't know if she coined the term or not, but for me it strikes the right note of competence within a strictly limited range.
Dan
| | | | Artyom Timeyev
Russian Federation
Local time: 13:13
Russian to English
+ ...
Quote: "It's not intelligence. It is about pattern recognition and statistics."
I have a friend who is a psychiatrist and he has been reading deeply into theories of mind and theories of schizophrenia. He's been reading about it for years. He is Russian, but has impeccable English, since he also a reanimatologist and loves to keep himself on the leading edge of knowledge.
He sent me a whole slew of links to research articles in peer-reviewed journals, and explained a l... See more Quote: "It's not intelligence. It is about pattern recognition and statistics."
I have a friend who is a psychiatrist and he has been reading deeply into theories of mind and theories of schizophrenia. He's been reading about it for years. He is Russian, but has impeccable English, since he also a reanimatologist and loves to keep himself on the leading edge of knowledge.
He sent me a whole slew of links to research articles in peer-reviewed journals, and explained a lot in a personal chat, about the current view on how delusions and hallucinations develop.
And you know what? Judging by all these advanced papers, written on the brain-network level, our brains have no "intelligence". It's about pattern recognition and statistics. Yes. Delusions arise in schizophrenia because your brain network starts seeing patterns where there are none. There is nothing "magical" in your mind, there is no homunculus there. It's a network for pattern recognition, and its workings are being deciphered fast, very fast. Almost every month there are news about this or that aspect of human neural network being discovered and understood. Ten years ago scientists could simultaneously record the activity of dozens of neurons, now it's into thousands. Ten years ago scientists could nudge the activity of a live brain network in one way (optogenetics), now it's in many ways, and this generates a lot of knowlegde.
And once you understand how a neural network works, you can recreate it in silicon. And it will translate, and compose poetry, and music, and jokes, and novels, and command complex manipulators.
[Edited at 2022-04-05 04:40 GMT] ▲ Collapse
| | |
|
|
|
Artyom Timeyev wrote:
Judging by all these advanced papers, written on the brain-network level, our brains have no "intelligence". It's about pattern recognition and statistics.
…
And once you understand how a neural network works, you can recreate it in silicon. And it will translate, and compose poetry, and music, and jokes, and novels, and command complex manipulators.
Perhaps. But there is no timeframe, is there?
What’s more important: these chips will never have the input means that we humans have. They can only process the prepared data derived from this input, that we supply. That’s a big disadvantage and probably it prevents real creativity and inventiveness.
| | | |
Artyom Timeyev wrote:
Quote: "It's not intelligence. It is about pattern recognition and statistics."
Regarding intelligence and creativity, I read this today. See what the robot itself thinks about this topic about halfway down the article:
Can she paint from imagination? "I like to paint what I see. You can paint from imagination, I guess, if you have an imagination. I have been seeing different things to humans as I do not have consciousness," she responded in stilted fashion.
This discussion about how far you can get with just pattern recognition and statistics is fine as far as it goes, but it avoids the "hard problem", which is why we humans have subjective experience of our own consciousness, whereas machines (presumably) do not.
| | | | Anna A. K.
Germany
Local time: 10:13
Member (2020)
English to German
| Localization... | Apr 5, 2022 |
Gerard de Noord wrote:
Don’t plan to study European languages to become a commercial translator. By the time you’ve finished, AI will have taken over. The same goes for interpreters: only top summits and peace negotiations will still require human translation.
Study European languages for other reasons but not to make a living as a commercial translator.
I’ve been in the business for more than twenty years, so just take my word for it.
Cheers,
Gerard
Translation is one thing, but a machine won't be able to do localization on a human level anytime soon. The profession itself won't go anywhere, but it will definitely change within the next years. We will all become proofreaders at some point.
| | | |
I don't agree. MT for translation is the same as a calculator for the stock market.
People still need to validate stuff and some complex work need a second eye and human work on it.
All the best,
| | |
|
|
|
Daryo
United Kingdom
Local time: 09:13
Serbian to English
+ ...
| Yes a bit, and mostly no. | Apr 13, 2022 |
Gerard de Noord wrote:
Don’t plan to study European languages to become a commercial translator. By the time you’ve finished, AI will have taken over. The same goes for interpreters: only top summits and peace negotiations will still require human translation.
Study European languages for other reasons but not to make a living as a commercial translator.
I’ve been in the business for more than twenty years, so just take my word for it.
Cheers,
Gerard
Yes, in 10 - 20 - 30 years it will not be the same as today.
AI in various forms will play an increasing role in translating - agree on that, there is no way that the trend could be reversed.
But it will NOT be the end for translators - not in any foreseeable future.
The number of translators might start decreasing and the entry requirements might start becoming higher and higher (who will need an amateur / chancer when MT can do it as badly, but far quicker and cheaper?), but I can't see translators going the way of the dodo anytime soon.
You could also argue that when it comes to AI not needing humans any longer to do translations, at that stage of its development AI might decide that it doesn't need humans at all ... Not a worry for today, or near future.
As for interpreting, you could theoretically combine speech recognition, processing of that input and speech generation, but to get to a stage when that would be something usable for anything more than some very approximative uncommitted chit-chat will take AGES. I don't see any reduced need for interpreters any time soon.
BTW "top summits and peace negotiations" are only a very small fraction of events where interpreters can most literally "make or break" a meeting. Penny-pinching on interpreters can produce far far worse results than penny-pinching on translations, plus the bad results are immediately visible. No one sane will be asking for "AI interpreting" for serious negotiations of any kind for a very long time from now.
Also, if and when in some distant future "AI interpreting" becomes a thing, then the "AI might decide that it doesn't need humans at all" part will definitely apply. So nothing to worry about.
OTOH having the possibility/capacity to use your knowledge of langages also for other things than translating / interpreting is a good idea for sure. If doing foreign trade, for example, being your own translator/interpreter will give you a definitive competitive edge.
| | | | Gerard de Noord
France
Local time: 10:13
Member (2003)
English to Dutch
+ ...
TOPIC STARTER | I hope you're right | Apr 13, 2022 |
Daryo wrote:
Yes, in 10 - 20 - 30 years it will not be the same as today.
AI in various forms will play an increasing role in translating - agree on that, there is no way that the trend could be reversed.
I hope you're right. I'll retire as a commercial translator before the end of this decade, and I hope the robots won’t catch up in my line of work.
After retirement I plan to translate lyrics, but to tell the truth I see AI is already making progress in that field too.
The thing is: AI gets better by the month. I’ve been using and testing DeepL since its inception and I state that DeepL’s English, French and German to Dutch translations have improved since last year and seem to keep improving continuously. And Dutch isn’t a priority on anybody’s list.
As with the ascent of CAT, let’s try to use our knowledge advantage in favour of our trade.
Cheers,
Gerard
[Edited at 2022-04-13 14:44 GMT]
| | | | | Depends, writing skills are a must | Apr 13, 2022 |
A vast portion of translation needs are covered by MT with barely any human supervision (news, online sales platforms etc.) just for the cost savings.
Then there is a whole segment of translation needs where agencies or in-house big companies use Custom MT models and have handpicked translators that review MT output. MT is kept in the first stage of translation for speed more that economical motivations.
Freelancers use MT to increase their speed or as help when they fe... See more A vast portion of translation needs are covered by MT with barely any human supervision (news, online sales platforms etc.) just for the cost savings.
Then there is a whole segment of translation needs where agencies or in-house big companies use Custom MT models and have handpicked translators that review MT output. MT is kept in the first stage of translation for speed more that economical motivations.
Freelancers use MT to increase their speed or as help when they feel fatigued.
But there's a situation where MT will not come anytime soon: where writing skills are important. MTs are biased and tend to stick to the same style (even with adapting layers such as glossaries etc.). This means that a translator with high quality writing skills won't loose market as long as they are discovered.
In fact translators with high writing skills tend to stay away from MT to not be steered towards the MT style.
So, my two cents would be: if you're looking to translation as a career, work hard on your writing skills.
[Edited at 2022-04-13 15:18 GMT] ▲ Collapse
| | | | | Agree with your points about good writing, but... | Apr 13, 2022 |
Philippe Locquet wrote:
Freelancers use MT to increase their speed or as help when they feel fatigued.
... I would've thought the worst possible time to use MT would be when you're tired.
The best thing to do when you're tired is have a rest.
| | |
|
|
|
Ice Scream wrote:
... I would've thought the worst possible time to use MT would be when you're tired.
The best thing to do when you're tired is have a rest.
You're 100% right!
What some do doesn't mean they should
Fo those who can, breaks are a must. Fresh air often means a fresh eye and that always help to improve quality.
| | | | Tom in London
United Kingdom
Local time: 09:13
Member (2008)
Italian to English
Philippe Locquet wrote:
..... work hard on your writing skills.
Indeed, and don't make basic mistakes such as "loose" instead of "lose" etc.
| | | | Gerard de Noord
France
Local time: 10:13
Member (2003)
English to Dutch
+ ...
TOPIC STARTER
Tom in London wrote: Philippe Locquet wrote:
..... work hard on your writing skills. Indeed, and don't make basic mistakes such as "loose" instead of "lose" etc.
How I'd love to correct your Italian pronunciation in public. I'm sure you suck at it, even after all those years. Your unnecessary comments are one of the main reasons I'll stop posting on this website.
[self-censured]
Gerard
| | | | | Pages in topic: < [1 2 3 4 5 6 7 8 9] > | There is no moderator assigned specifically to this forum. To report site rules violations or get help, please contact site staff » Consider the consequences of AI before you start a career in translation | CafeTran Espresso | You've never met a CAT tool this clever!
Translate faster & easier, using a sophisticated CAT tool built by a translator / developer.
Accept jobs from clients who use Trados, MemoQ, Wordfast & major CAT tools.
Download and start using CafeTran Espresso -- for free
Buy now! » |
| | Trados Studio 2022 Freelance | The leading translation software used by over 270,000 translators.
Designed with your feedback in mind, Trados Studio 2022 delivers an unrivalled, powerful desktop
and cloud solution, empowering you to work in the most efficient and cost-effective way.
More info » |
|
| | | | X Sign in to your ProZ.com account... | | | | | |
Login to reply/comment 
Translation news
