When resisting Russian aggression, Ukraine adopted a wide range of technologies based on artificial intelligence (AI). This gave the Ukrainian military certain advantages on the battlefield that would not have been available without AI.

Nevertheless, AI technologies not only offer new opportunities on the battlefield but also create social risks in terms of observing human rights and freedoms. Therefore, it is important to consider measures to be taken to prevent and solve the risks inherent to such a “military” AI application. 

Military AI application in Ukraine

“Nowadays, we are a kind of training area on the artificial intelligence application,” admitted Mykhailo Fedorov, the Minister of Digital Transformation, in one of the interviews. However, not only do we test Ukrainian developments in practice – but also well-known foreign companies use this “window of opportunity” for their technologies. However, many of them have previously been criticized due to security and human rights aspects (concerning the consequences of the application of those developments). 

In 2022, Ukraine announced cooperation with Clearview AI, an American company engaged in AI technologies for recognizing people using images from social networks and the Internet. It was reported that Clearview AI aid accelerated the process of gathering evidence to prosecute alleged war criminals by identifying persons in territories occupied both since 2014 and the full-scale invasion on February 24, 2022. This tool is also used to identify Ukrainian children forcibly taken from orphanages in the occupied territories, most of whom were adopted by Russian families. It is also used to detect Russian saboteurs and for some other intelligence and counterintelligence purposes. As of the end of 2022, the company facilitated finding 125,000 people in the territory of Ukraine.

“Following the military aggression of the Russian Federation against Ukraine, field units of the Ministry of Internal Affairs use the digital platform of the American Clearview AI to identify persons when documenting war crimes. The artificial intelligence created by the company allows law enforcement agencies to get identification information on persons using a special platform for investigations... Due to cooperation with American colleagues, we can effectively investigate the crimes of the Russian occupiers and identify criminals,” commented Leonid Tymchenko, the Deputy Minister of Internal Affairs of Ukraine, on the ongoing cooperation with the American company in April 2023.

In the same year, Ukraine adopted products from the American company Palantir, which is engaged in developing AI technologies for the military industry. The basic technology, Palantir Edge AI modular system, analyzes different data sets. AI-based software helps automate processes such as the takeoff and landing of an uncrewed aerial vehicle and capturing targets in combat missions. Alex Karp, CEO of Palantir, notes that their company is actively helping Ukraine in the military area and is responsible for “most targets” identified by AI in Ukraine. In addition, Palantir contributes to developing technologies used for the investigation of war crimes. According to Andriy Kostin, the Prosecutor General, the company, in partnership with Microsoft, has shown new ways of conducting investigations, particularly using data analysis and voice recognition tools.

Another American company, Primer, uses natural language processing technology in automated data analysis. The company identifies, transcribes, translates, and analyzes Russian voice messages using AI. Sean Gourley, CEO of Primer, claims the technology performs four functions: collecting and streaming audio, suppressing noise, and translating and highlighting key statements related to information about the battlefield. Data from unsecured Russian sources, such as social networks and radio channels, is also analyzed. Some other American companies, specifically Scale AI, shared similar technologies for image analysis with Ukraine.

However, according to military terminology, these AI technologies are of “dual purpose” and can be used in civilian life as well. Expansion of their application to new areas may entail risks to democratic rights and freedoms.

The main challenges of using military AI 

Even though the mentioned AI technologies ensure obvious positive results in the course of the military conflict with Russia, a significant challenge for their further use in Ukraine is the issue of ensuring human rights protection. For example, Clearview AI, which provides its services to Ukraine, is heavily criticized by the international community. The national regulators in Italy, Austria, and France have already imposed sanctions and multi-million fines on the company due to violations of the General Data Protection Regulation (GDPR). This literally froze its operations on the markets of EU member states. Therefore, given the Russian-Ukrainian war, Ukraine offers the company a chance to test and prove the efficiency of its developments in practice.

Representatives of Ukrainian civil society note that the further use of technologies, which are prohibited on the territory of some EU states, may slow down or even jeopardize the process of Ukraine's accession to the EU. Anna Liudva and Tetiana Avdeeva, experts of the Agency for Legislative Initiatives, indicate the following main risks of using Clearview AI technologies: dependence on the algorithm due to the non-human decision-making on a person's identification, challenges to confidentiality, and insufficient data protection from malicious interference in general. 

“The main claim against Clearview AI operations is that its tool denies a person's right to privacy. The technology could recognize anyone whose photo was at least once posted on any social network. Clearview AI is the most perfect tool of mass surveillance, the flawless embodiment of the so-called “surveillance capitalism”,” warned Nadiia Balovsiak, media expert, back in March, 2022.

In addition, AI technologies used even for purely military purposes cause concern among international organizations and experts. According to OpenDemocracy, an international media platform, criticism is primarily focused on the possible false target identification, bias, and excessive trust in the algorithm of biometric systems. The non-transparency of algorithms and the future possibility of non-compliance with the principle of personal data confidentiality also raise concerns. 

The issue of responsibility for the negative consequences of the system application remains unresolved. Oleksiy Kostenko, Dmytro Zhuravlev, and others draw attention to insufficient physical and ethical control over AI development and use. They believe that identifying malicious data is one of the biggest challenges to using AI for military purposes. According to Alex Karp, CEO of Palantir, “there are huge ethical issues on the battlefield”. “If you use an algorithm to generate a military decision and it goes wrong, who's responsible?” he says.

The above-mentioned factors pose risks for the further use of technologies similar to Clearview AI, Palantir, and other systems in Ukraine. Human rights protection related to using AI technologies remains a major concern for Ukrainian regulators in the wartime context. Vitaliy Honcharuk, Head of the Expert Committee on the Development of Artificial Intelligence in 2020–2022, notes that the war can be won “with really developed AI technologies, developed infrastructure and qualified engineering and management personnel.” At the same time, he believes that an important regulation element is to distinct between two scenarios of AI development – “peacetime” and “wartime” – and to develop appropriate rules for each.

Regulation potential

Ukraine is already taking the first steps toward AI regulation, particularly in terms of protecting democratic rights and freedoms when applying AI technologies. The Concept of Artificial Intelligence Development in Ukraine,approved in 2020, states that one of the main principles of AI development and implementation in Ukraine is “compliance with the rule of law, fundamental rights and freedoms of a man and a citizen, democratic values, as well as provision of relevant guarantees when using technologies in question.”

European integration processes have also brought the acute issue of AI regulation in Ukraine to light. In particular, in October 2023, the Ministry of Digital Transformation published the Artificial Intelligence Regulation Roadmap in Ukraine. The document does not envisage the development of standards and rules for AI use in certain industries – further AI regulation will be carried out under the “self-regulation” principle. At the same time, the Roadmap confirms the government's intentions to gradually implement the EU AI Act, which pays significant attention to human rights protection when implementing and using AI. 

According to the Roadmap, in the first stage of regulation, which lasts until 2025/2026, the government will try not to “overregulate” the AI industry while preparing businesses for legally binding regulation. At the same time, when discussing further AI regulation in the military sector, Oleksandr Bortniakov, Deputy Minister of Digital Transformation for IT Development, noted that during a full-scale war, AI regulation in projects related to military needs would not be addressed. Despite this, Ukrainian experts in the field of AI have already discussed the use of military AI after the war ends. However, these issues will require additional discussions and design of independent standards. 

“We have to use the zoning concept: what should AI regulation look like during war and after our victory? In wartimes, we plan to develop a legislative framework not to hinder the use of AI innovation and technologies to win. However, when the war ends, AI regulation should help Ukraine to further strengthen its democracy,” told Anna Bulakh, Head of AI Ethics and Partnerships at Respeecher and a member of the Committee on AI at the Ministry of Digital Affairs, Delo.UA.

The applicable legislation of Ukraine, particularly related to personal data protection, as well as law enforcement agencies and procedural codes, do not currently regulate AI use. Relying solely on self-regulation when using technologies developed by such companies as Clearview, Palantir, etc., may not be enough to eliminate all the risks inherent to the possible consequences of their application outside the battlefield.

Conclusions and recommendations

Artificial intelligence proved its efficiency during the Russian-Ukrainian war. At the same time, it is worth considering the risks to democratic rights and freedoms inherent to the development and application of AI technologies. These risks cannot be subject solely to sectoral “self-regulation”, especially when it involves the application of AI technologies beyond the area of fighting Russian aggression. Therefore, the main challenge is human rights protection and compliance with the principles the Ukrainian government must adhere to concerning AI applications.

In particular, the following measures allow for eliminating risks and ensuring fair and accountable use of AI technologies in the “peaceful” sector:

• Amending the applicable legislation: the key regulatory gaps related to the application of AI technologies for military purposes should be eliminated to ensure the protection of human rights and freedoms. For this purpose, the experts of the Digital Security Lab propose to amend the laws “On Personal Data Protection”, “On National Police”, and the Criminal and Criminal Procedure Codes.
• Ensuring the priority of regulation implementation related to the development and use of AI technologies in the military sector. The consequences of these technology applications are a vulnerable issue, which means that self-regulation alone cannot be reliable, as it cannot meet all the related challenges. The risk of interference with democratic rights and freedoms is the main argument in favor of proactive actions.
• Ensuring control over the use of these technologies. The results generated by the AI application must be verified by the responsible persons before final decision-making based on them. Moreover, law enforcement agencies' use of these technologies should be subject to judicial control, similar to that applied to covert investigative (detective) actions.
• Regulatory publicity of these technologies’ application. Ongoing discussions and debates involving experts and stakeholders from various industries are required to balance the interests of the authorities, civil society, and AI development companies.

The main task for Ukrainian regulation related to the application of military AI technologies should be ensuring relevant human rights protection that may be affected in the course of such application, especially if these technologies are used outside the battlefield. This should be the core idea when designing legal regulation of AI technologies, along with ensuring the industry's economic development.