 Hello everyone today today you get to spiffy me dressed up I had a lunch meeting and I came back and I figured I would stay well dressed for you guys and today I wanted to make a video that that kind of relates to that to my meeting that I had because Pretty much every time I go out and every time I talk to people whether they be translators or non translators But especially non translators they always ask me the question They always ask me Well, do you think now with Google translate or whatever machine translation? You know, do you think that you'll be out of a job? Do you think you know how much of a half-life? You know how how many more years do you have as a translator as a translation agency because sooner or later? They're gonna perfect this technology. There's no future for it And you know, I hear this with different variations But most people just ask me some people say well, you know there you have no future So what are you gonna do or stuff like that? And so I wanted to make a video that because every time they asked me this I kind of feel like I need a stock response and I wanted you know to print out a card where I can give the Response or maybe I'll just direct them to this video. This video is my response to all that and so here I wanted to go through it briefly or not so briefly and Show you what I think is happening and the reasons why by the way, I already did a video that talks a bit about machine translation but it sort of just more I kind of go through Google translate and show you some of the funny things you can do or some of the mistakes it makes and and And why that happens and so I'll link to that video in the description below but here I just wanted to talk more about About machine translation per se and about the effect this will have on translators on translation agencies in the coming years and And anyway might take on all that So will machine translation take over our jobs and will we become irrelevant because of it? The short answer is no so if that's enough for you then That's enough for you and you can watch another video, but if You want to know why I feel this way? Then let me get into it a bit more. So let me start off with first of all if you enter something into Actually, let me start off with Google When I say machine translate I mean stuff like Google translate or Bing translate now There are a bunch of different variations of this and and a bunch of different ones that then might be a bit more Specialized in certain languages or in certain fields or stuff like that like you'll find ones that are specialized Yeah, for certain languages and maybe only scan like financial documents to see what's been translated which way or maybe only official ones They're ones that have been issued by governments or stuff like that and and these are actually used quite often I'll get into that a bit later but Usually when people talk about this stuff they mean Google translate they mean these universal ones Google translate Bing translate their cistern Do they still have the babel fish thing? I don't even I don't know anyway You know all the all these other ones and that's what they mean. So that's The you know main thing I'm gonna be addressing here. So first of all You enter something in to say Google translate and you write something like I am going to the store chances are whatever language you put in It'll translate it correctly. That's a simple enough sentence. It stands by itself and You know, that's all you need to worry about. So that's fine The problems start to come when usually you notice the problems right away unless it's some Issue about a complicated sentence. You you usually notice it right away with metaphors similes Inuendos euphemisms stuff like that and and definitely humor humor is the worst usually and so here I'm talking about stuff like raining cats and dogs. I have an axe to grind Waiting for the other shoe to drop I have skeletons in the closet stuff like that These are all expressions euphemisms if you will that native speakers understand perfectly they have no problem with it and When you use it with a native speaker or even not a native speaker, they know what you're talking about They can make a judgment call. They know the gist of it and you can go on. However, these machine translations they work on an algorithm so many times these things are not very easy and So what do you have you'd say I have an axe to grind or it's raining cats and dogs Or I'm waiting for the other shoe to drop you put this and translate into another language if you put it by itself It might actually get the the expression correct because it scours all past translations and finds ones that are similar and Lo and behold it finds the right one. Unfortunately. We usually don't Use it this way. We often don't use it this way in fact for many of these the majority of the times We don't use them this way. We don't use them exactly like I'm waiting for the other shoe to drop or while waiting for the other shoe to drop It'll be something like, you know, someone says something you'd be like and the other shoe drops or you know, some variation thereof And it'll be something like you see this a lot with fiction You see this a lot with newspaper headlines and with people talking because language is always evolving and moving and people try To use it in slightly different ways, you know, you know, because that way they can grab your attention That way they can say something that's a bit of more original and so that's why it happens that way You'll see a headline that says This is a dumb example But let's say they come up with a law that says you're not allowed to bring your cat out in the rain, right in some neighborhood or some place And so the headline will say something from now on it's only raining dogs and then they'll they'll talk about it and at the end It'll say so from now on since no cats are going outside It won't be raining cats and dogs anymore, but it'll only be raining dogs something like that Anyway, everyone's gonna understand what it means despite it being kind of a dumb head dumb play on words But anyway, everyone's gonna understand the play on words. That's the important thing However an algorithm not necessarily it's gonna translate it as it's only raining dogs from now on and Which means nothing in pretty much any other language. I know of And there are other variations of this one is that very often we get them wrong The the other day I was having a conversation and this one guy was talking about someone He's like blah blah blah and he's too smart for his britches anyway And the la la and you know the conversation went on because everyone understood what he was trying to say He's like, you know, he's too smart for his britches. It was fine and it was clear except it's wrong that's not what the expression is it's supposed to be he's too big for his britches, right and So there again an algorithm. It won't get that right. It'll just translate it literally and it won't make any sense So and and just another example recently I was reading an article and it talked about an Akamite What did I say anyway it said it was a rather Akamite situation now it happened that two paragraphs earlier It referenced Occam's razor and it said you need to refer to Occam's razor in these situations Blah blah blah and then two paragraphs later. It says and this is somewhat Occamite So obviously Occamite, which is not a real word by the way I looked it up before doing this Occamite was referring to Occam's razor now It's perfectly fine to use it this way obviously, he invented the word if you will and just wanted to say Occamite to refer to Occam's razor and Anyone reading it will get it. You know, I understood it fine because it referred to Occam's razor right before so It in fact it only stuck with me because I had this video in my mind And so you know it clicked and and that's why it stuck with me But otherwise most native speakers won't even pay attention to that But once again the algorithm won't know that word because it's not a real word And so it won't compute if you will And there's so many other examples Obviously and you see this a lot also with a simple with nursery rhymes obviously with any type of poem with any type of song lyrics and then stuff like that Recently I was reading about how the cat on the mat if you enter that in Google translate It gets translated literally, but it's one of the first phrases any English speaker learns and it sticks because it rhymes That's why you learn it. It's one of those things the cat on the mat But and every language has its own equivalent, but it's not literally the cat on the mat. Anyway, fine So let's say Okay, we understand how all these sort of euphemisms metaphors similes, you know innuendos these these These expressions that have been that have entered our vocabulary It's a bit harder to use them with Google translator with any sort of machine translation So, okay, fine, but aside from that you might think machine translation follows a certain Algorithm and it's a very complicated algorithm and getting smarter every day So stuff that's more mathematical in language can probably be translated, right? So here I'm talking about technical texts I'm talking about financial texts legal texts stuff like that Except here you get into other issues now all these issues are pretty advanced Anyone who's used Google translate for more than a couple sentences In fact, you can go pick any paragraph in whatever language and translate it back to your native or whatever You're you're more comfortable with and read it through and you'll notice some glaring mistakes So Google still has to get through those, you know before it starts actually tackling all of these But anyway once it gets through those it has as I mentioned before all these expressions that it's gonna have to deal with And the expressions are constantly changing by the way and evolving so that makes it harder as well Also, when it comes to these more mathematical texts, there are a lot of other issues to deal with as well So I've seen because I deal with a lot of legal translations in this legal ease I see a lot of issues with a double negative English does not have double negative. You don't say you don't know nothing You say you don't know anything because two negatives make a positive. That's the feeling But this is quite artificial because actually they they tried to do this on purpose back at during the age of enlightenment or something It's when they decided that that should be the rule in English and most of the languages definitely western european. They don't have that So, you know, you don't know anything in in French In Italian, in Spanish, you don't know anything. So I mean all of these would be Double negative in all these languages And so you just put And you translate into English probably most of these machine translations will figure it out because it's not that hard And I mean once you do it once it's been seen a couple times in past translations It can get that part. However, what I see in these legal translations are extremely Weird convoluted sentences, you know, these lawyers trying to make them as convoluted as possible So you can't understand them so you need to hire more lawyers and keep paying them to figure them out I mean, you know, that's what they say, but uh, so okay, here's a sentence that I had just recently Uh, if I can find it here should the client here and after refer to inter alia The principal not have any indemnification with regards to or against any and all actions proceedings claims or anyway It you know, it's a sentence that no normal person would say and in there you have I mean not a double negative, but you have something that would be double negative in most other languages um, so This can be very dangerous and I have seen this in the past where if you try to use a machine translation It uh, it brings it out as two negatives because many times one of the negatives is here and the other one is here And so bring them out as two negatives, which means not only you have a mistake in translation But you end up saying the complete opposite of what you want to say if you want to say you're forbidden to do this In actuality, you'll end up saying you're not forbidden to do this which Can be very dangerous with a contract or legal, you know, any legal document. So That's something you need to be careful about another thing that you have with legal translations with financial stuff accounting Especially and whatnot is that in different countries? They are very very different So if I have a text if I have a text that's describing how things are in the states and they refer to You know miles that's a simple way to put it miles or degrees Fahrenheit or about certain legal issues or accounting standards or financial standards or something along those lines Then fine, you can translate it translate it correctly and then you're fine But if you're trying to write say a contract that's valid here and in another country Bolivia, china, australia, wherever it might be Then you can't just do that you need to convert it and you need to make sure it makes sense in Well in celsius or kilometers or in the legal You know for the laws of whatever country you're dealing with or for the accounting standards of whatever country or area you're dealing with Because there are many different accounting standards financial standards legal standards I did a video earlier about the different way you can write numbers and that's digits In all these in in switzerland in italy in the u.s. They're all different in china It's very different and you know and these are just with numbers So when it comes to actually writing things out things can be very different according to different laws and accounting standards, etc, etc google translate doesn't know this any of these machine translations don't know this by the way Many bilingual people don't know this. This is why you can't just hire someone who's bilingual to do these translations You need to hire someone who knows what they're doing so um that's Anyway, so that's one of the issues one of the big issues that you see and that you can come up across When you're dealing with all these, uh, you know legal more technical texts Let's say financial legal stuff like that also remember a lot of money is on the line for these And so they take them very seriously these a lot of these translations Basically what i'm driving at is that google translate is an algorithm and another argument i've heard is that well Okay, google has been so good at doing things like driverless cars you know they can definitely do translation and i might catch catch some flak for this but i mean Driverless cars in a way is almost easier than translation because you're not dealing with The way humans speak Driverless car in essence. I know it's a lot more complicated than this But basically you want to keep the person safe So you're like get from A to B and avoid any obstacle that's in the way, right? And so that is an algorithm I mean that's it's a very complicated one that i could never come up with And and obviously it needs to be tweaked and it's getting better and better all the time But you're basically avoiding the obstacles and Trying to get from A to B safely in one piece, but language translation Is not just that algorithm. It is that judgment call It's that it's that innate understanding that we need to have and i don't want to sound like oh We humans are so artistic that the machines can't capture because machines can capture a whole lot But it's that judgment call Where you get the gist of the situation everything that's happened. That's very difficult Just as it is now Google translate translates sentence to sentence and almost all in fact all the machine translations i've seen do that At maximum they'll refer to a sentence right before but anything that's too paragraphs before like the augments razor example I made before Um, you know, they're not going to catch that because there's too much other information in the middle and you know It's not sure should we refer to that or should we refer to something else? So i mean it's uh Anyway, it can it can be an issue That it needs to be it's not an issue of making the algorithm better or smarter It's an issue of getting a whole new algorithm a whole new type of ai if you will um And that is an issue and i think until they come up with this new type of ai That they won't be able to substitute real translators Let me give an example I think i've mentioned this example before but it's a very good example of this precisely and i got it from Steve Pinker by the way There's an expression that if i'm speaking to another native speaker, they'll have no problem understanding I'll say something like oh You know time flies like the wind and they were like, yep, it sure does Everyone understands that and it's fine and you move on with the conversation, right? And you're reminiscing you'll be like oh time flies like the wind and that's fine Okay, and you think okay, what's is that's not very hard in terms of uh in terms of an expression However, there are many different ways that an algorithm could read that Time flies like the wind Well, there's a joke that says time flies like the wind fruit flies like bananas And that could be a way to read it time flies could be a type of fly and they like the wind and uh fruit flies like bananas and so time flies like the wind so That could be another way to translate it. Here's another way Time flies. Here's a stopwatch time flies like the wind, okay So the wind is going and then you have flies going time them I guess in relation to the wind there's something like that, right? Um, and here's yet another way to see it a fourth way You can say i'm giving you a stopwatch. I'm giving the wind a stopwatch You time flies like the wind times flies time flies like the wind Now any english speaker is good is listening to this and saying okay, but those are stupid No one thinks that way. No, but they're grammatically 100 correct So any algorithm is going to see them that way any algorithm You know is it's that's the issue with any algorithm Let's say and so for this specific example Maybe you know it can figure it out because it's a pretty common one But it's just an example of something we as humans don't think about because we can make that judgment call right away but computers with their algorithms do have to process it and they have to make sense of it and very often that can be hard Especially when someone's giving their new take on uh on an expression and Time flies like the wind it can be time flies like an arrow time flies, you know like anything else Or it can be or the days fly like the wind or I mean, you know You're going to find different variations thereof or when they just get it wrong like the example I mentioned of someone who's too smart for their britches But then carries on the conversation as if nothing happened because everyone figured it out. So it's fine So anyway, that's just my argument if you will against trying to use this machine translation willy-nilly trying to use google translate willy-nilly. There's another expression willy-nilly in fact, you know, yeah try to enter that see if that translates into whatever language because You understood what I meant even if you don't know the expression willy-nilly you got the gist of what I was saying but Willy-nilly isn't the most common expression. I guess it it kind of is anyway, you know try with machine translation See what happens, but Here's what I get into. I'm not actually against machine translation I am against people thinking it will take over, you know or dropped or people thinking they can use that instead of translation I'm not against it. I think actually it's very important and there's certain situations in which it's very useful one is One is when there are emergencies you take if there's a famine or a war or a disaster in some place like Syria or Haiti or whatever it might be and you need You know, you need to communicate in a different language with someone and you need to ask for water For supplies for blankets stuff like that and you list them out, of course Use google translate because you're not going to use euphemisms. You're not going to use kind of weird innuendos They're new expressions or whatever it might be No, you're just going to list things that you need and then and the person receiving it by the way will understand like even If the translation is a bit off. They'll figure out what you mean You say you need bags bags could mean bags under your eyes in english And maybe they receive the wrong translation of bags under your eyes But they'll realize. Oh, no, it must mean bags because it's uh, anyway So when you have when you have an emergency, of course, you use it You're not going to try to wait for someone to do the correct translation or anything like that because time is of the essence The other situation in which I think it's used Correctly and well, um and very useful is when people who know what they're doing are using it People who know the ins and outs of the translation The mistakes that can be made especially the mistakes with machine translation Then it can be very useful and these experts by the way are translators and So someone like me I do a lot of translations in financial and legal in the financial and legal fields in italian, so I'll use websites like lingue and uh and What else is there there's reversal stuff like that they operate under the same idea they scour Other texts and see how things have been translated in the past, but They're uh, at least as far as I know what they do is they scour other financial or eu government You know official texts that can be so these are official ones that you can count on and you can refer back to So usually you don't translate whole paragraphs obviously you just translate a word or an expression and uh, and then it'll show The original and the translation that was issued by the eu or by whatever bank or something like that So if there are issues you can refer back to it or you can pick and choose you can say well I'll choose the official government one and not the bank one because I want to use the official government translation anyway For people who know what they're doing. This can be a great help So a lot of people say they don't want to use you know In fact people boast is say I use no machine translation at all. I'm like well. Why not it's it's a help You know if you know what you're doing it can be a tremendous help It can be very dangerous for someone who has no idea what they're doing and just figures They can enter something and just use that as a translation and so that's kind of what I'm against. So, uh, yeah How long have I been speaking quite a long time haven't I okay? Anyway I'll wind it all up. So hopefully this makes sense to you Hopefully here you grasp you have a better grasp an idea of how translation work and how Basically machine translation Is still behind and I don't think we'll get up to that level even if they perfect the algorithm because it needs a completely different algorithm One that has a sort of judgment call that captures the quote-unquote gist of the situation Which we humans do innately but machines thus far cannot do they can fake it pretty well But they can't do it because they don't you know, they think in terms of algorithms and in terms of programming and not in terms of Judgment calls like we do Uh, which means they're very very good at some things that we can never hope to be as good as them at but for other things We still have that advantage. Um, anyway, so I hope you found this useful and uh, yeah, I hope you can refer to this Especially at least some of these arguments, especially when people start asking oh is translation Is a field that's uh, not gonna exist anymore? Or is it uh, you know, is is it gonna die out or something like that? And the reason And actually I don't mind people saying it that much because I feel like the more people feel that way The fewer people will enter the translation world and you know, that's less competition if you will So if they if they want to think that way, then that's fine On the other hand if they think that way, then they're going to start using Google translate instead of a translator And that's going to be bad for them. So, you know, we'll see Anyway, that's it for now. I hope you found this useful Hope you appreciated the suit and I'll talk to you in the next video. Thanks. Bye Oh, and also don't forget to Subscribe if you want more videos like this Don't forget to click the like button if you do find this useful because then I can know what's useful and what is It gives me a better idea and Yeah, that's it. I'll see you in the next video. Thanks. Bye