Artificial intelligence will not save the Polish economy

It may turn out that instead of bringing success, artificial intelligence will plunge us into peripherality. The development of the IT sector will not automatically entail a digital transformation in other sectors of the Polish economy. Our advantage, i.e. entrepreneurship and great engineers, is not relevant. Polish deficiencies stand out, however: low level of digital skills among employees and managers, lack of capital to invest in innovative solutions, shortcomings in management.

Creating solutions in artificial intelligence and their implementation in other sectors of the economy are two different things, economically speaking. It may seem a truism, but hearing (true) stories about outstanding Polish programmers, we seem to forget that, silently hoping for an economic leap of the Polish tiger straight into the digital paradise.

Yes, Polish startups are successful, and our engineers are widely praised. Although the vast majority sell their products abroad. In an interview with Krzysztof Mazur, Agata Wierzbicka, the co-founder of Krakow’s Base, discussed this: „We never had nor planned a sales department in Poland”. Base was no exception; such behaviour is the norm because it is simply a rational business decision. This is confirmed by the Map of Polish AI, according to which the majority of revenues in Polish startups in this field are generated by foreign customers.

Therefore, hoping for a leap of the entire economy, not just the IT sector, we must not only ponder whether we want more AI Made in Poland, but also whether we care about artificial intelligence implemented in the Polish economy. These are two separate issues that largely require different actions.

Economically speaking, the first element is to build supply, the next – to build demand. The challenge will be greater in case of the latter since the starting point leaves much to be desired.

Demand: too little funding, too little research, too few specialists

Let’s start, however, with the development of the ICT sector. If we want more startups creating innovative solutions based on artificial intelligence and other advanced algorithms, we need two things: investment in scientific research and highly educated specialists – programmers, mathematicians, engineers.

We must immediately invest a lot more in AI research in Poland. Unfortunately, we will not be able to compete in the global arena by allocating around 1% of GDP to research. This constant element of diagnosing the situation of Polish sciences is even more relevant in an industry where universities worldwide face a drain of minds to tech companies that not only offer high salaries, but also the possibility of pursuing one’s own research unhindered (e.g. by grants or didactic duties).

Although we may find some glorious exceptions, the number of groups conducting AI research at a global level is scandalously low in Poland. Without building a critical mass – establishing a coordination centre, much like the Łukasiewicz Research Network, and becoming part of the European initiative of the ELLIS institute network – we will not be able to keep up with global progress. And without that, we won’t educate more world-class scientists and programmers at our universities.

According to the Digital Economy and Society Index, prepared by the European Commission, the percentage share of ICT specialists in the Polish economy is around 2.7%. That’s fewer than in Scandinavian countries (over 6%) or the European average (3.7%), which gives us the last, 20th place in the EU. However, this could be explained by our overall lag. After all, there are few statistics in which we would surpass Scandinavia or the EU average.

Something else bothers me, however. The authors of the summary of the Digital Sustainability Forum workshop cycle have noted that the percentage of ICT specialists in Poland has not changed over the last 10 years, while in Scandinavia it has increased by approx. 20%, in Germany or France by 70%, and doubled in Bulgaria.

The percentage share of IT sector employees in the economy isn’t changing at all. The number of IT students is decreasing (an impact of the demographic decline), and the demand for specialists is growing. Compared to other countries, we not so much stand in place, but rather go backwards.

Supply: why the poor assimilability of technology in Poland?

What do we have to do to implement more AI-based technologies in various branches of the Polish economy? The number of scientific papers or the number of specialists in deep learning is essentially irrelevant. A high level of medium digital competences is important, especially among the management; knowledge about running a digital transformation process, readiness to recreate an institution’s or enterprise’s processes in the digital world. Finally, which must not be forgotten, what it takes is capital.

Here I don’t have good news. Polish companies are slow to introduce technological innovations. Only 6% of Polish small and medium-sized businesses use data analytics, which places us second to last in the European Union.

The average Polish employee has poor digital competence when compared to their colleagues from other countries. According to the Digital Economy and Society Index, 46% of the population in Poland have basic digital skills, which places us at the honourable 24th place in the EU. This problem is well illustrated by the graph in the SpotData report:

The situation among mid-level Polish managers is no better. Here I will take the liberty of relying on anecdotal „evidence” – conclusions from conversations with a dozen or so startups at the turn of the last few weeks.

Managers in businesses must sift through a plethora of new technologies offered to facilitate their work, to choose those that can actually introduce better quality, and implement these innovations skillfully. Neither of these two steps is simple, and there is also the absolutely fundamental issue of funding the – usually expensive – programming projects and the fact that the IT department would often prefer not to take chances. Every new system in the company is an annoyance, if only for security reasons.

It comes as no surprise that, according to The Global Competitiveness Report, published by the World Economic Forum, we occupy a remote 68th place in terms of ICT absorption, while landing 38th in the entire ranking.

Conscientious process documentation vs Polish reckless soul

Supposedly, up-to-date and well-kept documentation of processes in companies or administration can be found in the same place as the Amber Room. Usually, if such a thing exists at all, it is a forgotten pile of sheets in a binder somewhere, labelled „Audit 2015”.

Nowhere in the world is step-by-step documentation of processes considered a priority. However, the Polish private sector and administration are particularly not very fond of it. Many startup entrepreneurs in the business process automation sector acknowledge that. A well-developed process seems to stifle the Polish reckless nature and our innate improvisation skills. A definite exception is the sector of shared services and outsourcing centres (BPO/SSO). After all, these processes had to be „virtualized” before moving the department to Poland and hiring new people on site.

Lack of structured processes and the scarcity of their documentation are another of Poland’s challenges. On the one hand, this means that Polish companies of all sizes will be losing their competitive edge against global rivals over time. Furthermore, with better preparation of processes, they will be able to implement AI-based technologies cheaper and more successfully.

On the other hand, the outsourcing sector may suffer. While the fears that along with rising salaries, service centres will start moving to India have proven unjustified, a scenario where they would move to the Amazon’s vast server rooms seems quite real. The shared service and outsourcing sector, hiring nearly 300,000 people in Poland, maybe largely virtualized.

The key to success is the regular employee

Yes, there are industries where employment may soon decline sharply. However, in the current market, a proven and loyal employee is usually a valuable asset. Companies automating business processes, robotic logistics centres or implementing algorithms in the industry are confirming the scenario of the fear of automation passing quickly.

Their products are often advertised as tools to replace the work of ten people with, for example, only one. Most often, however, the remaining nine employees are successively transferred either to more satisfying, less routine work (e.g., to respond to customers’ non-obvious questions, while the algorithm handles the simplest ones) or to other departments in the company.

After a department manager announces that a new technological solution would be introduced, employees often go pale with fear. After a few weeks of experimenting, however, it turns out that they are beginning to appreciate getting rid of the most difficult parts of their work and see their own efforts becoming more effective.

The implementation’s success, as well as the ability to raise one’s own competences to be able to work in another department, or take care of handling more complex tasks, depend on digital competences. It doesn’t take the ability to write code – a large part of the solutions implemented nowadays allows building software through a graphical interface, in the form of „blocks”. It’s more about understanding computers: the ability to solve simple technological problems, multiplying Excel cells and formatting tables in Word or verifying sources of information from the Web.

And how does it work in practice? On the one hand, there are people in the administration who would ask cringingly if they could have a website with an „@” in the address. In companies, on the other hand, some people prefer to run mathematical operations on a calculator, and only enter the result into Excel.

Let’s move on. Recently, at the Impact’19 conference, I heard an interesting question on the implementation of 'virtual assistants’ to help doctors diagnose. It went like this: „How is this supposed to be done if 30% of doctors in the (as far as I remember) Mazovia province are in retirement age?”

Conclusion: let’s stop being delusional

Undoubtedly, the IT sector in Poland is in for a time of prosperity. Yes, we need to educate more programmers and analysts – medium- and high-level alike. Nevertheless, the salaries in this sector will attract even those who will have to acquire their knowledge of the free market, for lack of places at state universities.

The hope that the sector will „trigger” a revolution in the entire Polish economy is futile. Polish startups primarily sell abroad. This is a rational business decision – euros and dollars allow paying salaries to valued employees; it is easier to find customers with the right budget and ones who are prepared to implement the technology. In Poland, even where the budget is not a problem in itself, there is often no will, preparation or skills necessary to assess risk and make decisions.

This, however, means that instead of using AI in the Polish economy, we will return to the periphery: some jobs will be virtualized, technologically backward Polish companies will lose their competitive edge. Struggling with their weak basic digital skills, Poles will have to overcome a high barrier to be able to change trade or simply improve their competences. The lack of technology in a company means lower competitiveness; the lack of digital skills means difficulty in changing trade.

In the Artificial Intelligence Quotient, we show examples of Polish companies creating AI-based solutions in sectors as diverse as finance, agriculture and the chemical industry. Our partners from the Digital Poland Foundation and Microsoft, in turn, diagnose barriers in the Polish economy.

Many of the issues raised here also appeared in the Assumptions for the AI strategy in Poland published by the Ministry of Digitization. We are on the eve of issuing the final version of this concept. The Future Industry Platform Foundation has also been launched. The only question is whether this strategy and other activities aimed at developing technology in Poland will receive proper attention, including political, and sufficient funding. Without it, in a dozen or so years it may turn out that the current semi-peripherality was quite a good situation.

Polish version is available here.