Originally the article was published @ European Commission – European AI Alliance.

 

How to develop an AI Government Strategy for Emerging Country  

I’m an entrepreneur, founder and CEO of a company, working in the field of AR and autonomous navigation for robotics in Palo Alto, California from 2013.

While my business is based in the USA, originally I’m from Ukraine and still have close ties with my native country.

I have several AI R&D offices in Ukraine, starting from 2014 I organize annual and the largest AI conference in Eastern Europe – EECVC.

In 2019, together with a team of proactive tech enthusiasts, we initiated(pro bono) the establishment of the Expert Committee on AI Development.

The idea was to develop AI policies and to put the country on the map of the most advanced countries in the field of artificial intelligence and machine learning.

Our motivation was very simple – if not us, then who? We had a clear understanding that this domain should be managed and controlled by professionals, otherwise it could be headed by people with no experience, no vision and no understanding of AI. In this case, final strategic documents could become not executable(The National AI strategy of Ukraine has been already approved by The Cabinet Of Ministers on December 2, 2020).

In this article I will share my experience, including the fight against internal bureaucracy, balancing on the edge of national interests and other interesting conclusions that can be useful for my colleagues in other emerging countries to develop their own efficient AI strategies for their countries.

The team is a key

The further you are from Silicon Valley, the fewer people speak your language and share the same practical experience and values as you do.

I think the key to the success of quick development of Ukrainian AI strategy was people.

The core team consisted of people I selected based on the following criteria:

1. Practical up-to-date experience in AI;
2. Global and/or international track record and global visibility;

Building a team according to these criteria saved us a lot of time, allowed us to resolve all misunderstandings and protect the development process against external attacks from the “old system”, which could lead to a delay or even suspension of development of Ukrainian AI strategy.

Organize the process correctly

At the first stage of working on a strategic document, the entire process should be kept as secret as possible.

The Ukrainian AI strategy was reviewed by 30 internal experts, 200 external experts and more than 30 government bodies. The process of approval took several months and was reviewed by 24 ministries (3 people from each ministry), which means that more than 400 people were involved.

It’s an enormous amount of daily communications and people involved. Don’t think that Excel spreadsheets will help you, use advanced communication channels for your team and tracking systems for any external communications.

Bureaucracy issues

In developing countries, old bureaucratic systems are trying to block any prospective documents. Without any exceptions.

Our main strategy was to keep the document and group of people in a stealth mode, avoid any conflicts with a bureaucratic machine until the document will be ready for a public discussion, otherwise we would be drowned in meaningless meetings and meetings.

Once we released the document for public discussion, we left no chance for the old system and time to mimicry except one – accept us as leaders in policymaking and join the process of public discussion of the AI strategy playing by our rules, not theirs.

Old bureaucracy in emerging countries is like a vampire – it is scared and avoids publicity as the light of the day.

Avoid cargo cult institutions

On our path of AI Strategy development, we found several “AI” institutions, representing cargo cult.

Such institutions have no successful track record for the last 30-40 years and they present a rather cargo cult mindset. I highly recommend to avoid such institutions and cooperate with ‘healthy’ government institutions only.

Otherwise, the development process of AI Strategy could be significantly delayed because of internal strategic discussions regarding focusing on “General AI development” and other strange topics, which looks absurd for any reasonable person with practical experience in AI.

Focusing vs copy-paste of USA/China AI strategies

When you are developing an AI strategy, you are trying to learn from leading countries – the USA and China.

These countries invested enormous resources in AI in the last 10-15 years. But since you are developing the AI strategy for an emerging country, you need to adapt the best practices for limited financial resources and human capital. At the same time, you need to be ambitious in your strategy and goals since it’s about your country’s competition in the global arena, not about developing ‘another strategic document’.

We decided to prioritise human capital development (e.g. how to use the government education system more efficiently for work force in AI) and focus on several areas of AI, where the country can achieve a leading global position. Otherwise, with limited resources, all these documents will be useless declarations.

Balance of national interests vs external technology licensing

Only a few countries in the world can compete in AI in all areas of implementation. Emerging countries need to find the right balance between licensing of external AI solutions and the development of their own.

This is especially important in the defense sector and other areas related to national security, where weapons become more and more autonomous. The question of who develops and controls their ‘brains’ is crucial.

Protect the process against any external influence

During the process of Strategy development, it’s important to identify people that are representing only national interests, not external interests. You need to identify people with any potential conflict of interests and exclude them from the process immediately.

Conclusion

When we initiated this process in 2019 together with the Ministry of Digital Transofrmation(Mr. Fedorov, Mr. Bornyakov and Ms. Ionan), I thought it would take us 2 years to develop a national strategic document or even it would be blocked by third parties. Surprisingly, it took us 10 months to have it done.

It’s a very promising sign when we see that emerging countries with old-school government apparatus can develop and adopt a very up-to-date strategic document in such arising area as artificial intelligence.

For sure, the adoption of this document does not yet mean its implementation.

Therefore, I plan to write about my experience of strategy practical implementation in all spheres of the state apparatus in the following articles, which may be interesting for my colleagues developing AI in other emerging countries.

Vitaliy Goncharuk

Tech Entrepreneur

The Head(pro-bono) of The Expert Committee on AI Development of Ukraine (The Ministry of Digital Transformation of Ukraine) A-14