Silver Kelk: “AI pushes us to redefine (im)possibilities in public innovation”

The Estonian Centre for International Development (ESTDEV) and Enabel have co-organised a discussion on AI as an accelerator for innovation in emerging economies.

Silver Kelk, head of business and innovation development for Mainor AS and Ülemiste City will represent Estonian AI innovations in Paris. We sat down with him to discuss how emerging markets can use AI and the role of public-private partnerships in developing AI solutions.

1. In the field of AI, what does Estonia have to offer emerging economies?

Silver Kelk (SK): Estonia still has all its well-known digital strengths, explicitly in the e-government and Smart Cities domain, but now you’ll hardly find any Estonian IT company or digital transformation consultancy that hasn’t adopted AI in their technologies or concept. In other words, Estonia keeps offering practical solutions that make sense from a digitalisation and innovation perspective, solving real-world challenges, but now finding extra leverage from justified and appropriate AI adoption in the mix.

Perhaps the most notable AI development in Estonia, that other countries and cities should monitor closely, is the Bürokratt, an interoperable network of chatbots on the websites of public authorities that allows people to obtain information from these authorities through a chat window. While both the world of AI and the public authorities’ massive websites and laws have been like unexplainable black boxes to the public, then in the combination of these two, we should have a smart personal assistant to our digital state, right in our pocket.

There are also many AI-assisted solutions coming from Estonian startups. In Ülemiste City, we have adopted large-scale traffic and mobility monitoring systems by Fyma and building automation solutions by R8 Technologies and tested several autonomous vehicles like Auvetech and Clevon.  I think these are great examples where AI is helping to solve real challenges and has provided an edge for these innovations.

Besides the AI solutions themselves, Estonia can also offer emerging economies our expertise on how to regulate new technologies so that it won’t hinder innovation capacity. It seems to me that the situation is similar to the last digitalisation wave where Estonia managed to gain many advantages by bold and flexible initiatives. But these weren’t just technological advancements that helped to build the e-Estonia concept, they were just as much the regulatory innovations and swift changes in public services operation that enabled these technical solutions to serve the market needs. In the age of AI, such a balance between regulations and innovation could be even harder to find, but I believe that the Estonian decentralised approach, PPP [public-private partnership] model, specialised sandboxes and regulation hacking accelerators (e.g. Accelerate Estonia) can provide smart paths for navigating securely through AI innovation.

Silver Kelk

2. What is the role of public-private partnerships in supporting AI innovations?

SK: I’ve always liked one realisation from “Regulatory Hacking” by Evan Burfield. In the case of public sector PPP innovation projects, when the project is successful, the technology company gets all the praise, when the project fails, the officials get all the blame. The risk-reward ratio is rarely balanced, and it often means that the answer to new technologies or operational methods is “It can’t be done.”

In fact, usually, it can be done, and it is already in the works in some private-sector companies. When governments and cities realize that in the right conditions, most of the innovation will come from the bottom-up and start in the private sector, the smartest governments create a suitable growth environment for these innovations. We see many examples in Estonia, where private-sector-led initiatives have been “institutionalised” over the years, growing from early innovators to long-term frontrunners and ecosystem builders, impacting much wider domains than their own economic interests. Ülemiste City developer Mainor is also a good example here, being one of the leading advocates for private companies’ data strategies, data collaboration and open data economy creation—understanding that our lead and example can positively affect and engage our growing community of 500 companies and 16 000 people in the district and even beyond.

But in general, the best thing that any public body could do is to avoid the “can’t be done” automatic response. With an open and curious mindset, there will always be some “how” that takes the process forward. The right balance will be found in open discussion, experimentation and compromise.

Silver Kelk, head of business and innovation development for Mainor AS and Ülemiste City

From the public sector perspective, it’s necessary to spot and support such private-sector initiatives, not necessarily rush to compete or regulate the space before even knowing what needs to be controlled or managed.  In the case of experimental technologies (like AI or blockchain), specialised sandboxes and exemption policies can help a lot. In Estonia, we have also built up a good innovation support system that addresses mostly horizontal enablers (e.g. digitalisation, data interoperability, sustainability, etc.) instead of focusing on vertical domains. Such a system usually gives more flexibility to the public sector to adjust its programmes and support systems in the wake of new opportunities and emerging initiatives. But in general, the best thing that any public body could do is to avoid the “can’t be done” automatic response. With an open and curious mindset, there will always be some “how” that takes the process forward. The right balance will be found in open discussion, experimentation and compromise.

The tricky question, especially for the European public sector, could be partnerships with international tech giants. Previously the cloud, now the AI era, has only intensified the digital sovereignty and data protection questions. I don’t have a perfect answer to this challenge. But again, the Estonian experience has shown that you can be successful both in developing your own solutions and in partnering with global market-leading companies. Here again, the right balance with regulations and policies, with an approach to creating open technological ecosystems would probably help to mitigate most of the risks and boost the domestic economy, as well. The general philosophy should be “trust, but have means to control.”

3.  Let’s look at the importance of data. Can we talk about using AI for emerging economies if there is insufficient reliable data?

SK: Sure, it depends on the applications. There are endless use cases for AI-assisted technologies in specific domains where the amount of data is not so critical, or the required data for training the algorithms are not so location-specific. But usually, data reliability is an even bigger issue. The true potential of AI can thus be unlocked with continuing digitalisation and strong data strategies. Today, such initiatives should be planned and developed with an open and collaborative mindset, extending data interoperability capabilities across sectors and borders. Here, the European Union and leading digital nations like Estonia have already done most of the job, not just with the technology but also with replicable concepts, guidelines and training.

Actually, there’s always room for improvement in any country regarding data collection and data quality. I’m keen to see how Europe involves more private-sector organisations’ and even individuals’ data in such larger regulated and secured data pools. It’s another huge challenge, mostly from acceptance and legislative perspectives, but I remain hopeful and believe that developments in data altruism, data economy and international private-sector data collaboration will give a huge push to innovation, including AI, in the coming years.

4. What about Europe? Why should European organisations aim for collaboration with emerging markets on data and AI? Where do you recommend they start?

SK: As an e-Estonian, I’ve always liked to play with the idea “What would you do differently today, if you could do it all over again?”. It’s no secret that the internationally-used Estonian IT sector can build a much more modern e-government concept today than the one we have for ourselves. Without the legacy burden, you could focus more on the present and future opportunities, skip some former stages, and perhaps build up a much more resilient and future-proof digital system—and why not with AI as a central component? I think this is also a lucrative opportunity for Estonian companies.

It’s not just the technology providers that keep developing. Also, Estonian digital transformation consultancies have gained a much wider perspective from all over the world in the past decade. For example, in Mainor, we have been building our Future City concept model with many academic and technology partners. It’s a comprehensive and very practical concept that includes the general Smart City vision, strategic domains and their applicable strategies, with specific technological solutions and a data-driven metrics system. On every occasion when we share this with any other Smart City, despite their progress stage or specific location, it’s fulfilling to see how our 15 years of practical experience in developing Ülemiste City and the alignment with future trends and technologies enable us to create a better tomorrow beyond our own countries and cities.

The conference can be followed online on March 26th.