How can the cold calculus of AI and machine learning create deeper human connection? What if it were possible to harness vast amounts of data about how people learn, and feed it back into a system which makes learning 10 times faster? This is precisely what Lingvist seeks to do for language learning.
AI and machine learning are 2017's buzzwords, but ask what they mean in their current state — or how they impact the day to day life of the average human being — and the answers are often vague or uninspiring. Ott Jalakas, Chief Operating Officer of Lingvist, shares his view on the present and future challenges for education — and how technology can accelerate the speed at which it’s possible to learn.
Why are people turning to technology to educate themselves?
For one, technology makes education so accessible – technology allows people to have access to education at a relatively low cost and independent of their location. This is significant because it means that the same educational material can be scaled to a higher number of locations and a larger number of people who in turn can benefit from the educational material.
Secondly, technology can optimise and enhance the learning process. Lingvist is a good example of this. Our primary aim is speeding up the language acquisition process itself, and also minimising the time actually spent on the language acquisition process. Through computer-directed exercises we can practice the most important vocabulary, at the optimal time and in the optimal way for us to remember it. The most exciting part is that through the use of machine learning, we can customise each course to build upon the learner’s previous knowledge, skill set, and learning tendencies.
To what degree can an app help someone learn a language?
The degree to which an app could help someone may include reducing the costs you spend on language acquiring: making learning faster; giving access to learning material, driving your motivation, the cost effectiveness compared to other methods. Personally, I believe that in the coming 3-5 years all learning material will be available in an electronic format. Where right now the standard is that we have hardcopy texts as the default and that material in its electronic form is a supporting format, I think this model will be reversed, with electronic channels coming first and hardcopy textbooks being derivatives or support materials.
Is AI the future of teaching yourself new skills?
Lingvist uses narrow AI and machine learning to drive course customisation for learners on our platform. We believe that Lingvist is one of the very first applications where AI techniques and machine learning methods are applied in real life — where you can truly benefit from them on a very palpable, everyday level. There’s a lot of buzz about AI right now; everyone is talking about it and experts agree that AI is the future, but it’s still very theoretical in many ways.
What is the benefit of using an app which uses machine learning algorithms to learn a language, as opposed to using an app which sources its content from a community of people, as seen with many other language learning platforms?
Different language learning apps cater to different learning styles and goals. Some learners have a more leisurely approach to language learning, whilst others do it more as a hobby. In such cases, having a community centric learning approach might be the best. At Lingvist, we focus on speeding up language acquisition and minimising the time it takes to learn a language. Hence, if you are aspiring towards scoring a language learning goal within a limited amount of time, Lingvist is the tool for you.
How will increasing automation through the use of AI affect businesses and what will the benefits be?
Everyone who reads or listens to the popular media in 2017 can see that AI is all around us and poised to expand exponentially. Many companies are using it as a buzzword in their promotional materials, including when describing their values, identifying it as central to their business strategy, and so on. In many cases it is still label-only and hasn’t yet been filled with real AI content and/or hardcore backend substance. Still, I think there’s no doubt that we are aiming towards using AI, deep learning methodologies, and other related methods more and more in the coming years, and there’s no going back on this. It’s irreversibly going in that direction — so much so that one really cannot ignore it.
The market is saturated currently with apps claiming to teach people a language in hardly any time – what makes Lingvist different?
It’s true that the focus in language learning has been shifting from gamification methodologies and peer-to-peer tools to having efficient and pragmatic goals. This in itself reflects the sort of efficiency, time optimisation and content relevance which sets us apart from several of our competitors who have been shifting focus away from peer-to-peer tools or gamification tools, or just traditional language learning methodologies. Even shifting focus like that is a hard thing to do if your core algorithms and backend are built with an aim to prioritise some other kind of aim, and that means legacy architecture needs to be changed. What differentiates Lingvist is that from day one we have built our systems, algorithms, our backend, and our study material with the primary aim of speeding up language acquisition. Our legacy and the core basis of our software and product philosophy has been very much aligned with the aim of minimising learning time. This gives us a huge advantage. We don’t need to change legacy architecture, we don’t need to rebuild software, and we don’t need to change the goals of the organisation as we have these critical things already in place.
What are the biggest challenges for language learning apps like Lingvist at the moment?
At the moment we are focusing on user retention. In order to have an impact in that area, you need to have 2 things in place. Firstly, the tool that you’re offering must be good and deliver value, and secondly there needs to be enough users who are actually using your tool. The overall value proposition to the world is the tool and its benefits in terms of time to the users. This is the ultimate utility that you deliver to the world, and our challenges are derived from these two aspects: We want to improve Lingvist as a tool; continually adding and building exercises must always be built in a way that their main aim is still fast language acquisition. And the second is to have impact on as many people as possible. So far we haven’t really focused on go-to-market strategies and user acquisition because we have been focused on building the best engineering team we can, and on machine learning and code. Now as we believe we are getting market-ready, we need to start building a user acquisition machine in a similar way to how we have been building the backend engineering team so far.
Why should people learn a different language and how can technology help motivate people to learn a language?
Technology improves the accessibility of language learning. Firstly, direct accessibility through mobile devices and computers gives you access to language learning tools like Lingvist. Secondly, Lingvist minimises the time needed to learn a language. This makes language learning more accessible to everyone because time is our most scarce resource. If I’m already busy with my job, with my family and hobbies, I have only half an hour every day to do something extra. If I choose language learning to be this “extra” then I can spend my additional 30 minutes on it, but then I want to use these 30 minutes as efficiently as possible. In this way, by enabling fast language learning and fast language acquisition, we actually motivate people because we demonstrate that “OK, learning a new language is actually doable” and it’s accessible in the rather compressed time-frame you can allocate. If you use a book, and you have say 20 minutes every day, it may take several years to learn a language. If you have 20 minutes every day and you use Lingvist then language learning is much shorter, and that’s why it’s much more motivating for someone to take up language studies within this substrate.
What’s the future of the Estonian start-up scene?
Estonian startups are excellent case studies of startups built in a way that, from day one, have a focus on the global market. The Estonian domestic market is so small that you cannot really build a company that focuses on the domestic market. So from day one you need to have the global market as an aim in your first slide deck of what you’re doing. We have several positive examples — Skype, Transferwise, Pipedrive, Lingvist — so people see that there are real people, people you quite often know personally, who have actually made it. This is encouraging – if they have made it, you can make it as well. Those who found startups who have made it or had successful exits often choose to reinvest their earnings into new initiatives in the local ecosystem to improve and build their businesses. I think that these two features are the cornerstones of the success of the Estonian startup scene, and will allow it to continue to grow and flourish in a sustainable way.
Image Credit: Lingvist