Me versus “Her”: MT versus AI

Samantha: I want to learn everything about everything.

Theodore: I love the way you look at the world.

(Turns out she has the mind of a translator!)

Here I am, as promised, or threatened, some days ago on Twitter. On Sunday, I partly watched Her. It has been a really interesting moment, with a movie that, respecting my original thoughts, was not really a great way to convey a message that, on the contrary, was interesting. In any case, contrary to all that I could have imagined, it touched me and made me think about different issues.

When I first heard about this movie, I just thought it was another “let’s try and make a huge philosophical point about human beings by using a really complicated surrounding tech world that prevents 80% of the audience from understanding the movie at all, let alone from receiving the message” kind of movie (OK, I should totally start finding shorter names for movies genres!) So, with this idea, I was not too much interested in watching it. After having started, I can admit that it has some value and I should maybe finish it (it was not my decision not to watch it fully,) but I still think that it can fit in my genre.

In its defence, I have to say that this movie is powerful, and I am not only talking about the fact that you totally plunge into it almost without noticing what they have done to Joaquin Phoenix, even if, after, you feel an unstoppable desire to watch again Walk the line to wash away the Groucho Marx impression on your retina. No, seriously, I have gone through this one reliving in my mind different movies that, one way or another, tried to talk about the loneliness and depersonification of our modern society:

We begin with Theodore/Phoenix estranged from the rest of the world, living all his life through the vocal inputs that he sends to his phone. It reminded me immediately of Wall-E and the society that the little robot finds on the starship Axiom: a group of obese human beings who travel on floating chairs and whose needs are satisfied by the main computer, without any desire for them to move; in fact, they don’t even realise what is around them, or that there are other humans on their sides.

Then, Theodore buys Samantha/Johansson, and in front of my eyes the screen creates a complicated mix of You’ve got mail and Bicentennial man. Theodore is alone, recently separated but too scared to sign the papers for the divorce, because he has not fully recovered from those feelings. He is not really able or ready to flirt with another woman, or is not capable of seeing what he has around, because he is too lost in his self-pitying, introverted world in which he is only comfortable when dealing with his devices and expressing every side of himself through voice commands and electronic inputs. At the same time, he decides to try a new operating system and buys Samantha. She is an OS that is built with conscience and learns fast, creating a superior being that is only lacking a material body. If at the beginning it may seem silly that they feel something so strong for each other, very soon we see the OS becoming a complete person (with all the doubts and incoherencies that the movie cannot resolve), and very little difference exists between Theodore’s and Samantha’s relationship and any other taking place long distance.

Letting on a side the sentimental part, the main subject that I wanted to approach in this post were the similarities between the doubts about artificial intelligence, AI, and the ones about machine translation, or MT. MT is the translation delivered by a machine, with no human intervention, and must not be mistaken with computer assisted translation, or CAT. Examples of CAT are SDL Trados Studio, CatsCradle, and Déjà Vu, softwares that allow the translator to save time and optimise the time of a translation: Studio creates translation memories and makes them easily accessible for future works; CatsCradle helps translating websites without having to directly modify the HTML; Déjà Vu is a software similar to Studio, and combines TMs with example-based techniques to give the best matches to the translator.

On a complete different plan, MT just offers its final translation based on different techniques and without any human intervention. The most famous example is Google translate, but it is not the only one. There are different techniques that can be used for a MT, and the main ones are: rule-based MT, and statistical MT. The statistical MT uses bilingual corpora to translate similar texts, and the bigger the corpora, the better the results. It is the main technique used by Google translate. The rule-based MT, instead, mixes different techniques, using the entries of the dictionaries and interlingual representation of the source text to obtain a target text.

Now, translators often, not to say always, criticise Google translate, making it the cause of all the problems that we face in the translation market, but that is not correct. Of course, some ambiguous instruction booklets that I read and that said “not to touch the kids” instead of “keep out of the reach of children” are almost certainly Google’s creations, but that doesn’t mean that Google’s results are always horrible, and that we cannot use it to have a quick idea of what our friend published on Facebook in a language we don’t know. We just need to be careful and not trust literally all the results to avoid embarrassing situations, like when I wrote that my dad loved boiled fish, literally “likes it to die” in Italian, and my reader asked me why my dad would die if he would eat boiled fish!

This brings me back to the original idea for this post, which is the interaction with humans and computers. Of course, the movie poses problems more complicated than the simple human intervention in a machine translation but, in some ways, raises similar questions, like what will we do when the machines will be advanced enough not to need us anymore?, and will that day ever arrive?

As for the movie, I see the man/OS relationship as a metaphor for relationships facing some sort of separation, either physical or cultural, but also for the constantly growing human dependence from tech devices. Talking about MT, the debate about the possibility of it overtaking human translators has been on for decades, and the danger is no bigger now than it was when I started my studies. So far, computers have made translators’ lives easier and their work more productive with hugely accessible information, time-saving tools and simplified communication. As far as I am concerned, I think it will take a while before computers really threaten our lives, physically and professionally, and by that I mean that for now I still am the one able to throw the other to the wall, and not vice versa!

Leave a comment

1 Comment

  1. At first, when yesterday as I saw the title of this post retweeted by someone I follow, I was pretty upset, because I was thinking that it was just a bad timing to see it published and crossing my path just a couple of days after I published my post about MT and AI. Also, still sure of what I wrote, I dived into the article.
    I have to admit that it is a really nice, interesting post, and that actually confirms my descriptions, and clarifies them, because I didn’t spend too much time on describing the different kinds of MT. Therefore, if you want to know a little bit more about it, please follow the link:
    There is something else that I wanted to add on the main post, but then I got distracted by the rest and I left it out. The image of Theodore walking in the middle of the crowd, or standing in the metro, listening to his device and giving it orders about what to do, like: “Play melancholic songs,” reminded me of two moments of my past life. The first is not a specific moment, but something that has characterised some months of my first experience in London. I used to walk to work, because it was really close, and to practice interpreting on the way, because I had a very important impending interpretation to perform. To do so, I had my .mp4 player, and I used to spend the 15 minutes to reach my job listening to speeches and simultaneously interpreting them. Sometimes, being really into it, I wouldn’t pay attention to my surroundings. Other times, though, being really self-conscious about it, I could catch people’s glances and frowned faces seeing me talking to myself in a foreign language. To avoid feeling like a crazy woman walking in London, and talking about the peacekeeping operations, or the EU fishing policy, or the latest invention to save energy, I would often hold the cable of my headphones pretending that there was a microphone and I was talking to someone on the phone.
    The second example is a memory from my previous, and first, visit to New York: I was walking down 5th Avenue with my host, and we reached Central Park. Tired and thirsty, we had a pretzel and some refreshments, and we just sat down in the square facing the park. A very pretty, well dressed, elegant woman was walking down the street in front of us, carrying her really expensive shopping from the boutiques around us. She was angrily talking out loud, and we both asked ourselves what was going on before we could find the strength to ask the other. At the end, the person with me asked: “Is she on the phone, or just crazy talking alone? I was kind of relieved by the question, because I was thinking the same: I didn’t see any headphone, and I was absolutely convinced that she was talking with herself. No one seemed to care, because it is also an important thing of big metropolis: no one care about other’s people doing. The only thing I know is that, until today, we are not 100% sure, but pretty much close to that, and it gave a lot of sadness to see her walking alone and talking to herself. Maybe we just misinterpreted the scene, but it stayed in my mind as the symbol of a lonely, self-centred, computer based society that could bring us to Theodore’s level very soon.



Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: