What does Croatia need to put itself on the AI map?
A brief guide to complement policy
Aco Momcilovic, EMBA. FutureHR, National AI Capital
Interview with:
Terence Tse, Nexus FrontierTech, Visiting Professor on Cotrugli Business School
About discussion he had with CroAI members:
Hajdi Cenan, Airt.ai and
Marijana Sarolic Robic
In your experience, what is happening with AI on the European level?
For the past several years, the European Commission has been producing numerous policy documents with the single aim of promoting the development and deployment of artificial intelligence (AI) based on “European values and rules” among the member states. This initiative is undoubtedly and urgently needed as AI technologies are evolving at neck-breaking speed. Furthermore, the ramifications and implementations that this technology has in shaping our economies and societies in the coming decades are expected to be vast.
Yet, despite such unquestionably commendable effort in leading a high-level discussion and making a long list of recommendations on the role of AI should play in the industry policy of individual member states, there remains seemingly a lack of concrete ideas as to how individual governments can build up their own AI industries.
Does this matter to Croatia, and what should we do?
This matters a lot to a country like Croatia — one that does has got neither a large industrial nor domestic customer base. Countries as such must become agile and stay at the technological forefront to be better able to compete on a global basis. Governments of these countries must devise a simple and proper policy framework to allow and empower companies and communities to have more flexibility in the development and deployment of AI technologies. Drawing on our experience from working in the AI field, this short brief intends to offer insights into some of the current developments in this field, which in turn will hopefully be useful in informing the right national policy for the Croatian government to build up its AI capabilities.
How are bigger countries reacting?
Governments in larger economies in Europe such as France, Germany and the UK have all made AI a key industry focus. The result is that they have got an earlier start on the development and deployment of AI. Contrast this to where Croatia is on the subject. At the time of writing, the government is expected to have completed the final version of the national strategy for AI. It is among a small handful of countries out of the 31 EU member states that have yet to establish an official policy.
As can be expected, the private sector does not wait. More and more Croatian AI companies have been emerging in the past years. As an illustration, the Croatian AI Association (CroAI), created in December 2019, has now got 71 companies as a member, with an additional 50 supporting members that are either individuals or academic institutions. For a fledging industry, these may be huge figures. But such a growth trajectory is not insignificant as it clearly demonstrates the country has got the right mindset, talent, and resources to thrive in this area. While this is a good start, for AI companies to thrive, it is necessary to have a more coherent collective effort.
But we have a number of companies that are developing AI projects?
A telling tale is where the very few AI companies in the country ended up: they often do business abroad and consequently applying their expertise overseas and hiring staff from the countries in which the transactions take place. There is no doubt that export is a good source of income for Croatia. However, selling services abroad can neither promote domestic consumption and investments of AI nor the AI capabilities of the country effectively. Indeed, taking a broader economic view, Croatia has got a good training and education system that produces a lot of qualified engineers and developers. Yet, the country has been suffering chronically for not being able to create enough jobs created to absorb the talent produced.
What is Croatia’s potential and what should be done to achieve it?
In addition to the EU’s efforts in technological development via policy framework, we would like to offer a few ideas that can potentially set a “bottom-up” agenda and guideline to achieve the goal.
Have conversations, not merely top-down. With the right policy and actions, Croatia has got the potential to become a regional AI leader, if not beyond. While the EU has been eager to push forward the development of AI, its focus is often restricted to establishing the right environment and regulations, which is not necessarily offering the quickest route. Indeed, the discussions of ideas should be broadened beyond the policy framework mindset. For instance, the ministry developing the national strategy should involve all interested parties and not just National Digital Economy Council’s Working Group and the few private companies.
What are the traps we need to avoid in creating AI value?
Focus on the real side of AI. It is very easy to have fallen into the following misconception trap about AI: it is all about AI models involving only AI researchers and scientists. In business reality, however, deploying this technology into company operations involve a much wider set of competence and job roles. The most critical success factor in the development of AI is not the strength (e.g. accuracy) of the models — the “brain”. Instead, it is the overall IT system in which the model is only one of the many components, with the so-called AI Operations (AIOps) creating and supporting it. AIOps involve system developers and engineers — and not data scientists — who are skilled at embedding AI models into a well-oiled production environment, which, in turn, must work flawlessly with companies existing IT infrastructure. Put differently, the way to genuinely extract the value of AI is not the model development but rather having a robust overall technology system.
Viewing from this vantage point, the essence of AI conversations should be extended to pushing for developments in the capabilities to run AIOps. In doing so, not only can Croatia develop more comprehensive and pragmatic AI offerings; it can also capitalize on the full potential of the developers and engineers churned out by the education system. A case in point is the London-based AI company Nexus FrontierTech that hires some 80 system developers and engineers in Hanoi, Vietnam, essentially helping the country to upgrade the skillsets of the workforce there.
Is it only about developers and engineers?
Think about the business side of technologies. It is important to note that not all activities related to AI involves people with technical backgrounds. Typically, a business that is considering the deployment of AI in its own activities will have to consider issues and requirements related to data, skills, cost, integration, and stakeholder education. Additionally, just as important, if not more, is the ability to strategize, plan for and executive AI projects. Having a strong AI project management ability is mission-critical as it will feed into solving many problems across the board. Much of the competencies to overcome these challenges are far more about business and less about technologies or technical skills. Taking this view means that there is plenty of scope and opportunities — as well as work to do — for Croatia to enhance its AI capabilities and therefore general competitiveness.
What would be your advice to the Croatian Government?
Instead of a 400-page strategy that seeks to obtain EU funding, there should merely be a set of strategic guidelines that can steer and promote the developments of AI start-ups as well as the wide-spread application of the technology. This should be developed by taking into account not just the AI technology itself but also the required business competencies for realistically deploying and developing the capabilities. If Croatia were to take on a broader view of AI — much wider than the focus of the EU — it will have a far better chance of thriving and excelling in this up and coming technology sector and put the country on the AI map.
