19-May-2019

The big press release of the season (possibly of the year as well, even though we’re less than halfway through) is without a doubt the European Commission’s Ethical Guidelines for Trustworthy AI1 (published in Apr-2019). Most developed states have their own AI strategies. That list includes my home country, Poland, which released an extensive AI strategy document in Nov-2018. The document (which I may review separately on another occasion) is not strategy proper, it’s more of a summary of key considerations in formulating a full-blown strategy.

It is likely that the lack of an explicit strategy (and lack of explicit commitment to funding) led Polish academic community to publish their own Manifesto for Polish AI.

Poland may not seem like an obvious location to launch an AI (or any other tech) startup, but anecdotal evidence states to the contrary. There are pools of funds (including state subsidies and grants) available to tech entrepreneurs and a relatively low number of entrepreneurs competing for those funds. The most popular explanations for this are the state’s attempts to stimulate digital economy and to stem (maybe even reverse) the pervasive brain drain, which started when Poland joined the EU in 2004.

Entrepreneurship is one thing, but research is something different altogether – at least in Poland. In the absence of home-grown innovation leaders the like of Amazon or Deep Mind, research is almost entirely confined to academic institutions. Reliant entirely on state funding, Polish academia has always been underfunded: in 2018 Polish GERD (gross domestic expenditure on research and development) was 1.03%, compared to EU’s average of 2.07%2 3. In all fairness, the growth of GERD in Poland has been rapid (from 0.56% in 2007) compared to the EU (1.77% in 2007), but the current expenditure is still barely a half of EU’s average (not to mention global outliers: Israel and South Korea both 4.2%, Sweden 3.25%, Japan 3.14%4).

Between flourishing tech entrepreneurship (18 companies on Deloitte’s FAST 50 Central and Eastern Europe list are Polish5 – though, arguably, none of them are cutting-edge technology powerhouses like Dark Trace or Onfido, both 2018 laureates from the FAST 50 UK list6), widely respected tech talent, nascent AI startup scene (deepsense.ai, sigmoidal, growbots) Polish academics clearly felt a little left out – or wanted to make sure they won’t be.

Consequently, in early 2019 Polish academics from Poznan University of Technology’s Institute of Computing Science published Manifesto for Polish AI7, which has since been signed by over 300 leading academics, researchers, and business leaders.

The manifesto is compact. Below is an abridged summary:

Considering that:

  • AI innovations can yield particularly high economic returns, as evidenced by the fact that AI is currently a primary focus of Venture Capital firms.
  • The significance of AI for economic growth, social development, and defence is recognised by world leaders.
  • Innovative economies worldwide are based on strong commitment to science and tertiary education. The world’s most innovative countries such as South Korea, Israel, Scandinavia, USA, and increasingly China happen to also contribute the highest proportion of their GDP’s to R&D and education.
  • Polish academia has significant potential in the field of AI.
  • A barrier for growth of innovative start-ups in Poland is not just lack of capital, but also, if not mainly, too low a number of start-ups themselves
  • Innovative start-ups worldwide are developed mostly within academic ecosystems.
  • There is an insufficient number of IT specialists in Poland, and even more so of AI specialists.

We are calling on the decision makers to develop a strategy for growth of key branches of R&D and tertiary education (in particular AI), and to take decisive actions to fulfil that strategy:

  • Severalfold increase in expenditure on primary research and implementation of AI to reach parity with the most innovative countries.
  • Growth and integration of R&D teams working in the field of AI and improvement of their collaboration with the industry.
  • Ensuring PhD’s and other academics doing AI research receive grants and remuneration comparable to those in the (AI) industry.
  • Additional funding for opening new AI university courses and broadening the intake into the existing ones.
  • Funding for entrepreneurship centres (based on the Martin Trust Center for MIT Entrepreneurship) on Polish university campuses, offering students and employees entrepreneurship education, seed funding, co-working spaces etc.

Funds committed to the above will be a great investment, which will pay for itself many times over in the form of greater innovation, a higher number of AI experts on the job market, and in effect faster economic and social growth, and better defence.

There is nothing controversial in the manifesto – in fact, probably most academics worldwide would sign a copy with “Poland” being replaced with their respective country name. It may be a little idealistic/ unrealistic (reaching R&D parity as % of GDP with global leaders is… ambitious), but that doesn’t diminish its merit. I for one would be ecstatic to see Poland committing 3% or 4% of GDP to R&D. Separately, it’s really nice to see that a compelling argument can be presented on literally 1 page. Less is almost always more.