The War in Ukraine and AI Regulation: Some Controversial Takeaways
Originally published at the European AI Alliance.
Since the beginning of the war, I have often been asked what The Expert Committee on Artificial Intelligence is doing now. Are we still doing policy-making? Certainly not.
Since the committee is a social institution, the same thing happened to it as to other healthy structures during the war — it was urgently reorganized and focused on solving the main problem of the country. Some members of the committee are now in the front-line trenches; others volunteer, work for the authorities, advise the military on modern technologies (including AI), and directly implement projects.
Background
In December 2020, the Cabinet of Ministers of Ukraine approved the AI development concept and its implementation plan, preceded by a year of public discussion and significant changes, as civil-society organizations mainly defended human rights, privacy and ethics — not security or the use of AI in case of war.
It’s not that human rights, AI ethics and privacy aren’t a priority — they truly are #1. But among the experts trying to regulate AI there were many experts on ethics and privacy, and very few who practically understand how AI works and its limitations. As Head of the Committee, I often had to receive public accusations on the verge of speculation — “inhumanity”, “desire to subdue the world”, “chipization” — from people whose understanding of AI is more “magical” than realistic. With the beginning of the war, everything fell into place; some public opponents privately apologized.
What did the war show?
1. Russian agents
During the development of the document, many incomprehensible (unqualified) people tried to wedge themselves into the process — even a law firm from Russia sent us a proposal to develop the concept. These characters took a lot of time, offered idiotic or manipulative initiatives, and delayed the adoption of the document — mainly citing ethics and human rights. With the beginning of the war, many of them declared a public pro-russian position. These were russian agents whose goal was to reduce all normal initiatives to zero.
2. Military use of AI should not be removed but discussed separately
We tried to create an ideal document with no public complaints, and had to remove many things from the defense and cybersecurity sections. Looking back, that was a mistake. We should have designed the document for two scenarios — “peacetime” and “wartime” — so useful developments would apply in a “Plan B” without peacetime compromises.
3. Data-privacy and AI-regulation requirements should differ for small countries
Since 2022, the bulk of the work on Ukrainian AI systems now used for military purposes was taken over by the private sector — activists or private companies. Small and medium-sized countries don’t have resources to create state AI systems in advance for such crises; many systems were created without privacy requirements because the goal was to survive. If these activists strictly followed the laws, perhaps Ukraine would not be doing so well on the battlefield. Therefore, small countries shouldn’t blindly copy the legislation of large countries and should foresee “special regimes” — including what to do with personal data accumulated during the war after it ends. The ability of countries to quickly create AI systems used in war is the key to survival.
4. Don’t forget the country’s competitiveness
Activists often want to regulate the industry and protect everything as fully as possible. But as soon as I bring competition with other countries into the discussion — and the fact that the Western world can’t afford to lose the AI race to China and russia — opponents switch from dogmatism to searching for compromise and practical solutions.
Summary
The war can’t be won with piles of documents on the regulation of AI — it can only be won with really developed AI technologies, developed infrastructure and qualified engineering and management personnel. If Western countries want to compete with China and other potential adversaries, it is worth significantly revising the balance between regulation and direct investment in technology development, and removing players with a conflict of interest from the regulatory process.
Best regards,
Vitaliy Goncharuk
LinkedIn