Ensuring access to up-to-date and trustworthy information is critical in a crisis. In this interview at NordicBaltic.Tech, a partnership between PUBLIC Denmark and the Nordic Council of Ministers, we explore how AI-driven technologies are supporting citizens and governments. We spoke to Kaspars Kauliņš, International Business Development Director at Latvian Tilde, to get his perspective on chatbots and cross-border, multilingual communication during and beyond Covid-19.
Can you elaborate a bit on your role as Business Development Director, Head of Language Technology at Tilde and building a Covid-19 chatbot?
My role was to lead the project – manage the team and allocate resources, coordinate the relationship with our project partner Tet and other institutions involved as well as to take care of PR and dissemination of our Covidbot.
Can you describe some of the key insights and learnings from the build?
Early on, when the global pandemic hit our country we realised that people are desperate to find reliable and trustworthy information about the disease, its spread, symptoms and mitigation measures as well as recommendations and restrictions that have been imposed by international and national authorities. We decided to create a single source of information using our conversational AI platform and technological tools. A large part of our efforts initially went to identifying trustworthy sources of information covering different aspects and topics. We analysed official websites of relevant institutions and social media to identify most important topics and frequently asked questions. Special attention has to be paid to the responses and building the chatbot’s knowledge base. We used different tools and means – condensed textual answers, infographics, videos etc. Where longer responses were required, we provided links to the respective websites. As information kept changing, we had to establish a process on how to follow up on the recent developments and make respective changes to our Covidbot’s knowledge base and dialogue scenarios. Where possible, we took dynamically changing content from reliable third party databases i.e. global statistics on infected people by country, using API calls to the respective databases. Another insight is that people on our team were highly enthusiastic and dedicated to accomplishing our mission quickly as everyone had a sense of importance and of the usefulness of our work. And it proved to be so as we got many regional and local authorities and professional associations interested in deploying our chatbot on their websites.
And the challenges – where do you see the biggest obstacles when it comes to deployment nationally and in a Nordic-Baltic setting?
The key challenge to introducing new AI-empowered communication solutions is related to the digital skills and user habits of our citizens – people have just started to get used to these alternative communication channels and tools. As there is not much empirical experience regarding talking to chatbots when communicating with public institutions, a certain level of hesitance and even resistance from these authorities is natural. Another obstacle is technological – languages spoken by smaller communities (compared to English, Spanish, Chinese or other largely popular languages) have relatively weak technological support. This applies to all our Nordic and Baltic languages that are not only under-resourced in terms of digital tools and means, but also rather complex and morphologically rich, which is rather challenging for any language technology application e.g. speech recognition and synthesis, natural language understanding, machine translation etc.
Despite taking individual approaches to combating Covid-19, all countries face the same challenges. In what ways might technology offer solutions that transcend national boundaries?
This is a global crisis and we have to fight it jointly. AI-driven technologies are there to help us first of all with multilingual communication and information exchange across borders. By using, for instance, machine translation facilities, people will surely benefit from gaining access to information and experiences in other countries on how best to deal with this situation. Language technology fosters accessibility to the trustworthy information sources, including for people with special needs – hearing and sight impairments, paralysis, dyslexia etc.
How do you see the role of the Nordics and Baltics when it comes to technology, innovation and government technology adoption?
I believe that our countries can clearly have a leading role in early adaptation of different innovative technological solutions as our societies are open minded, well educated and IT savvy. We are in a very good position of being relatively small in size and having a very good infrastructure in place. And our public institutions are also open to innovating and deploying new technologies for the benefit of our countries and people. I can bring an example of my own country, Latvia, that was the first country in Europe to introduce the language technology platform www.hugo.lv for public use, making AI-empowered solutions for the Latvian language available to the people. Finland is successfully introducing their Artificial Intelligence statewide program AuroraAI (https://vm.fi/en/auroraai-en). And there is more in Sweden, Denmark and Norway. And I believe that we have a great opportunity to join our forces and efforts, open our technological achievements to each other and become at the very least European if not global leaders in the application and introduction of AI-enabled technologies.
What is the role of government in the realisation of innovative tech solutions that emerge at the moment?
I see the role of the government at least as two-fold. First, governments can support development of innovative solutions and evolution of state-of-the-art technologies by investing in research and development projects and activities as well as in the respective education programs. Second, governments can help with the digital transformation of our societies and advancement of digital skills and habits by introducing these technologies themselves – into their own organisational processes, in government-citizen communication, providing e-services etc.
Startups – and the solutions they have to offer – may be under threat from the economic impact imposed by Covid-19. What are the appropriate measures (by governments, investors, local communities) needed to keep innovation afloat?
This will be a difficult time for start-ups as this pandemic brings a severe economic downturn which in turn has an impact on investors and businesses. We see that governments are already working on and providing large scale rescue and financial support measures to stimulate the economy. I strongly believe that each crisis also brings new opportunities for innovation and new value creation. Therefore I believe that we will see new ideas and companies emerging and striving. And the role of the state is to foster such initiatives and help new entrepreneurs to get started.
Much of the ‘tech toolkit’ being assembled by governments is designed to address a crisis. What can governments do to address the long term?
In order to address challenges and provide the benefits of modern technologies, governments need to facilitate the development of technological ecosystems – building infrastructure and integration of different technological platforms, opening data for AI solution training and public use, and opening up to each other by sharing and integrating national solutions for the common good. We will remain in a global competition with the superpowers in the US and China, which is why it is of utmost importance that we join our efforts in building a solid scientific foundation and collaboration between academic institutions, research centres and businesses here in Europe and globally.
About Kaspars Kauliņš
Kaspars Kauliņš is a Business Development Director at Tilde, a European language technology innovator and service provider offering multilingual conversational AI, translation, localization and custom machine translation.