Will Artificial Intelligence Remove the Need for Human Intervention?

I interviewed Gianna Brachetti-Truskawa, who is not only a linguist but a Senior International SEO Manager in Cologne, Germany. Gianna works for Bold Ventures GmbH. We are really pleased Gianna is speaking at Search Elite and I wanted to ask her a few questions regarding languages and the introduction of automatic translation.

in”I am impressed with the number of languages you speak. How many years should you study a language for before translating into that second language? Some say one should only translate into their mother tongue.

  • The amount of time needed to become proficient in a new language depends on a lot of things, e.g.:
  • Whether or not you grew up bilingual
  • Whether the new language you are about to learn is in the same language family as your mother tongue
  • Your approach towards learning the language – e.g. by reading books, watching films or series with/without subtitles
  • How much time you spend with native speakers, travelling to or living in the target language culture (i.e. the country or region where the new language is one of the officially spoken languages)
  •  Your ambitions and skills. To some people (second-)language acquisition comes easily, to others it doesn’t.

Nevertheless simply being bilingual (or quite fluent in another language) does not make you a professional translator. There are certain techniques  and strategies translators have to learn (for more details, see Mona Baker’s “in other words“. Translation requires a lot of practise, too. Studying helps but it is through practise that you get really good and efficient.

So it’s hard to name a specific number of years. For instance, I started learning English at school starting from the age of nine until 20, and later studied translation and interpretation for English for another four years. That’s roughly 15 years, not counting the years I have been using English both at work as well as in my time off (we must not underestimate the power of practise). Still, without having been living and working in the UK for a while, it would have taken a lot more time and effort to feel as confident as I do today when I have to switch from speaking German to speaking English.

Today, anyone can work as a translator, without a diploma. This is both good and bad. On the one hand this enables talented people to do something they love without having to invest a lot of money and time to get a degree, on the other hand professional translators can have issues to market their services successfully because of bilinguals rushing into the market and working for just a tiny part of the price of a pro. Many bilinguals feel they can translate “naturally” but lack the strategic skills and, often, have not lived in the target language’s country (or not been there for a long time). A certificate does not make you a good translator, still it is strongly recommended to invest in proper education and practise to become a professional.

With this in mind, it is commonly recommended to only translate from the foreign language into the (primary) mother tongue. Translation is a complex task and requires thorough understanding of the target culture and market. Culture can manifest itself quite subtly in communication, especially when it comes to context. Therefore, even within a given language, say: English, the same words might have a different connotation in Britain as opposed to the United States or Australia.

In the UK, it is no longer compulsory to study a second language after the age of 14 which I feel is is a real shame. How important is it to learn a second or even third language? Do you use all of yours?

Native speakers of English could argue that with English being the lingua franca in most countries of the world, it would be a waste of time to master any other language. Yet I think it is crucial to start learning a new language at a young age, and to never stop learning.

However, learning a new language involves learning different cultural concepts as well (given that you want to master the language beyond simple conversations for travelling). Hence I think learning a second or third language always makes sense, as it makes you more flexible when expressing yourself, and more tolerant towards others. Plus you can help others more easily to learn your mother tongue, as you will know the main differences and difficulties.

With each language you learn, your brain learns new concepts of language, therefore: new ways of thinking. Even more so if the foreign language is from a different language family than your mother tongue or the first foreign language you learned. This can help your ability to look at any matter from another perspective. You might even become more proficient or creative in your own mother tongue, as you learn more about the differences between your mother tongue and the new target language. It can also help your creativity. Each language can express certain concepts that others can not, or not as efficiently. Trying to adapt or recreate grammatical concepts or phrases and metaphors from a foreign language into your mother tongue can lead to quite surprising, refreshingly new ways of self expression.

When it comes to my passion for languages, I am fond of grammatical systems. So some languages I learned in order to have some basic understanding of their thought concepts, while I intended to actually use or work in others. The degree of my language skills therefore largely depends on my initial goals – and the time I can devote to practise. Unfortunately, language skills are like a muscle: use them or lose them. So in fact, I do no longer speak all the languages I ever learned, which is quite a shame, but I still kept the basic understanding of their concepts. I rarely forget things related to grammar, but easily forget vocabulary if I do not use the language actively for a while. It’s always a struggle to stay fluent or improve with a language you already know when learning languages is not your primary job. Fortunately, it is still sitting somewhere in my brain – so if you find yourself having similar issues, rest assured that it is all going to come back to you once you start using the language actively again. It might feel odd and frustrating at first but you are going to do a lot better quite quickly.

In general I enjoy diving into an entirely new language much more as opposed to practising one of the languages I already know but I guess that is normal. I try to read as much as I can in any of the languages I learned, or invite Couchsurfers to use the language more actively.

What cases have you seen where AI is preferred over Human Translation in Localisation or have you seen more cases where human translation is needed instead?

So far I have only seen AI being used to create product descriptions in some online shops. I recently spoke to a person currently developing a tool for AI-driven content automation to understand how AI has to be trained to create ‘natural language’. Apparently there have been attempts to use AI to create football news; however it seem that this is not working so well yet. It does work well with product descriptions for some products, for instance in fashion. Currently this makes a lot of sense if you’ve got a large online shop with a lot of similar products (when it comes to structure and product type, such as smartphones), but is a lot harder to implement successfully when it comes to complex products, services, or news. I think we are going to see a lot of interesting development there within the next two years.

Content is going to be more often created or translated by AI. Search results, especially when done by voice search commands, are also going to be delivered by AI (which is already, partly, happening, see here. In many cases, human intervention is probably still going to be needed, however the way and the extent to which it is needed is subject to change.

Now, as a former translator, I can assure you my heart still bleeds whenever I hear or see someone preferring automated translation over natural, human translation. However, seen from another perspective, technological advancements could even serve as bridges to online language gaps. With English being by far the most commonly used language on the web, we are talking discrimination and a possible loss in diversity, as not all informational resources are available in all languages.

“(…) if you are a bilingual speaker of English and Zulu for example, there are clear advantages to using the English edition of Wikipedia with close to 5 million articles, over the Zulu edition with only 685. “

Source: Holly Young, The Guardian

 

 

 

As a consequence, people of a severely underrepresented language are running the risk to abandon their mother tongues because they learn that they have to use the web in another language (such as English, Chinese, or Spanish).

 

 

 

 

 

The importance and influence of language on the web is also stressed by the Guardian:

“In 2011 the UN declared access to the internet as a basic human right. It is clear however that access alone is not enough to put everyone on an equal digital footing. As the internet and social media become increasingly embedded in how we connect with and understand the world around us, so too does the language we use to access that experience.” Source: Holly Young, The Guardian

Translation could close the afore mentioned gap, and with AI-equipped translation, localisation becomes quicker accessible and more affordable, so it might spread more quickly (so does content creation in a different language). However, the attempts towards AI and language processing are currently mainly focused on what is seen as the most important (read: the languages with the biggest target audience), not the most underrepresented, languages.

What specific areas of AI will be growing in 2018 where it will reduce the need for as much intervention by humans?

When it comes to language on the web, I think there are three major developments where we are going to see improvements in 2018 or shortly thereafter:

  1. Voice search gradually becoming the major type of search people perform
  2. Availability of translated or automatically created content on websites and platforms like Facebook, Twitter, and the like
  3. Live interactions through AI, such as live-interpreting through voice-enabled devices, call services, or Augmented Reality built into apps.

At I/O this year, Google showed a few interesting new features worth mentioning – such as a recording of Google Assistant calling a hair salon to make an appointment and a new feature for Gmail, Smart Compose, aiming to help users to write emails with the help of suggested text blocks based on what you start to type. They’re even adding Augmented Reality Directions to Google Maps, so you can use your phone’s camera to navigate. The impact of Google’s AI equipped services is going to be huge, and happening a lot sooner than one might have thought.

In fact, Google could replace us sooner than we think. As Paul Miller puts it in his article “How close is Google to building a Paul?”:

“(…) Google can already write most of my text messages, and now it can help compose for me. (…) I can drive a car. Google can drive a car. I like to remind my nieces and nephews to be polite. Google can remind kids to be polite. Sometimes I remember to call my sister. Google knows when I’m most likely to call my sister.”

With all those exciting news, one might wonder:

What can a common SEO do to face the challenge of technological progress in the not-so-distant future?

I am not saying classic search is going to die, but it sure is going to change. Google are trying to make us actively SEARCH for something less often, by showing us what they determined we like right away. Even if we wanted to search, the way we are doing it currently and results being presented to us is going to change, in fact this has already started with voice search and virtual assistants.

 

 

Here are some things you could do to get ready for that future:

  1. Learn about product development/management: It’s getting more important to figure out what people are really fascinated by, or what they really need.
  2. Learn about technology: Try to understand the latest developments and safety issues, read Google (http://www.seobythesea.com/) or Facebook patents (http://stks.freshpatents.com/Facebook-Inc-nm1.php?archive=2018). You don’t have to become an AI developer, but you need to know more about its possibilities and its limitations.
  3. Learn about other marketing channels: Yes, the ones you have been neglecting so far. It is becoming more important to know how to assemble a proper strategy and find out the areas you can grow your marketing in case SEO is not providing sufficient or relevant traffic anymore. You might have heard about Growth Hacking – talk to people who have proven they’re good at it and learn from them.

Apart from language- or search-related developments, we see AI creeping into many different areas, such as radiologylawagriculturesexual health, and desire. And if you think of Japan as being mainly exploring and improving the field of smart sex dolls, you’re mistaken: they are also developing AI-equipped robots for nursery homes.

Instead of pushing forward further AI development, however, we ought to discuss ethics, e. g. what AI may do/must not do or how much we can allow AI to support us. Contrary to what we might expect, this is not being ignored by the big players in the tech industry. At F8 conference earlier this month, Facebook presented their approach to AI ethics: “Fairness Flow”, a system acting like an “ethical compass” when building AI by checking for bias in their algorithms.

Yet this is just one part of the issues that ought to be discussed. We have to find ways to cope with the inevitable shift within our societies, for instance when it comes to our careers: While the use of AI is going to create some new jobs, there will also be less ‘routine’ jobs for humans in the not-so-distant future. Instead of focusing on building new things as fast as possible, we need to find out how to provide all members of our societies with alternate stable sources of income and a new definition of what creates their feeling of “making sense”. (For further information, follow Aytekin Celik who is giving lectures about ethics and AI.)

There’s a 4th point I’d like to add to my previous list of suggestions for my fellow SEOs:

Learn “irrelevant” stuff. Care for those interests of yours that you have been neglecting in favour of your job. What are you fascinated by? Is there any entirely new topic you’d like to dive in? Give it a try and dedicate some time – it might prove helpful to get your mind of the classic SEO wagon for a while, while improving some new skills that might one day be useful. Either for creative ideas for a new product/service, or just for finding an alternative purpose when search as you knew it has changed substantially.

Thank you Gianna for your time, we can’t wait to see you at Search Elite on June 6th.

1,148 thoughts on “Will Artificial Intelligence Remove the Need for Human Intervention?