Russia’s AI setbacks will likely heighten its cyber aggression

With the weight of Western sanctions crippling parts of the Russian economy, the consensus seems to be that Moscow’s ambitions of being a major player in the development of machine learning, robotics, natural language processing and other artificial intelligence (AI) tools are functionally dead. The consequences of the war waged against Ukraine on Russia’s wealth, workforce and access to sophisticated imported products such as microprocessors used to operate everything from mobile devices to automobiles are immense.

Without capital, talent and a line on critical commodities and technologies, Russia will struggle to be competitive in everything from medical technology development to national security practice. This likely result of increasing isolation seems doubly assured with AI. Russia’s relatively weak fundamentals and strong competition from both China and the West virtually guarantee vast opportunity costs to Russia in years to come. This outcome might be seen as a positive development that will cede techno-strategic advantages to defense communities in North America, Europe and East Asia–those most concerned about Russia’s military capabilities and intentions.

There may be some truth in that assumption but to equate the recession of Russia’s ability to build cutting-edge AI with diminished likelihood of conflict centered on AI would be disingenuous. Practitioners and policymakers across the West would do well to recognize that the subversive character of AI design and deployment creates perverse incentives toward conflict even, almost especially for actors that do not lead the field. In no place is this coming uptick in conflict more likely to manifest than in cyberspace.

Why Russia’s AI efforts lag

In 2017, Vladimir Putin famously stated that “[a]rtificial intelligence is the future not only of Russia but of all of mankind” and that “[w]hoever becomes the leader in this sphere will become the ruler of the world.” Indeed, AI has come to sit at the heart of landmark initiatives established by the world’s leading powers to transform the conflicts, economies and societies of the future.

Mirroring efforts in Europe, a National Security Commission on Artificial Intelligence in the United States has encouraged immense increases in non-defense AI spending, widespread talent development and recruitment programming, and revamped acquisition vehicles to make harnessing new AI potential easier. India has become the leading adopter of AI tools among emerging economies. China has established unprecedented investment regimes tied to programs like the Belt and Road Initiative – and defined by the 2017 New Generation Artificial Intelligence Development Plan – designed to leapfrog perceived American technological superiorities.

Despite Putin’s statements, Russia’s AI efforts have lagged behind most initiatives in other countries. In 2014, barely two years after a breakthrough innovation of deep neural networks by a multinational group of researchers energized AI development, Russia’s buildout of new machine learning applications and other AI tools was already slowing significantly. Collaborations with cutting-edge projects in the West, China, India and elsewhere began to drop away following Putin’s annexation of Crimea and decision to embroil Eastern Ukraine in ongoing conflict.

State-sponsored companies and military-intelligence institutions in the Russian Federation have consistently been a leading source of novel AI technologies aimed at bolstering national security and strengthening mechanisms of population control. However, a slow leak of human capital and a complicated relationship with parts of the global economy that dominate critical high technology resources, such as graphics processing units (GPUs), have become a substantial obstacle for the country’s AI ambitions. Even if Moscow matched China’s or the United States’ levels of domestic AI investment, its fundamentals of innovation for the field simply haven’t been concrete for some time.

One major blow to Moscow’s AI ambitions is the dramatic acceleration of the brain drain that has plagued Russian high technology and scientific communities for years. Enticing researchers out of private industry and academia is perennially difficult for governments, but Russia has been even less capable in this regard than most, likely due to the unappealing culture and benefits of Putin’s military and paramilitary communities. Now, up to 70,000 tech workers that were otherwise minimal flight risks have fled the country. Many have ended up in former Soviet states and South Asia, and no small number have left positions tied to the Russian state’s focus on building out facial recognition, autonomous vehicles and surveillance capabilities.

Sanctions worsen the impact of this brain drain by cutting across the research and business relationships that Moscow has heavily supported in recent years. In the face of shaky fundamentals for domestic AI development, Putin’s government has emphasized collaborations across industry and academia with India and China. Publication and patent activity seem to suggest that this approach may have been paying off for Russia.

Despite producing a fraction of AI-linked patents (50 or ~5%) last year relative to the United States and China, Sino-Russian and Russo-Indian research teams collaborated on hundreds of research papers in the same period. Now, with such collaborations subject to international sanctions on Russian industry and citizens, this approach seems destined to flop. If anything, this dynamic virtually ensures the Russia will become an even more minor actor in economic relationships with Asian countries as the absence of American firms and the loss of any ability to lead collaborations removes the leverage that Russian industry might previously have had.

Russia’s use of AI in cyberspace still a concern

This lesser position in the global “AI arms race” has done little to quiet fears about Russian use of AI for conflict and interference. Moscow’s emphasis on “grey zone” tactics (mechanisms of state power that help shape favorable conditions without provoking escalation such as disinformation, cyber operations or the infamous “little green men” of Russia’s 2014 campaign in Crimea) feed such concerns. Combined with Putin’s adversarial rhetoric against the NATO alliance, these methods have produced understandable worry that any edge in AI could contribute to more strategic ambiguity more assertiveness in Western-Russian relations.

Liberal use of AI-generated deepfakes and other productions in the ongoing Russo-Ukraine war that push false narratives feed this fear. At the same time, a lack of sophisticated AI usage in military operations can been explained away in the same way a lack of major cyber operations was, namely that digital tools shape conditions around conflict but aren’t useful on the battlefield.

On the surface, it’s easy to see how recent events suggest advantage to Russia’s strategic competitors. Yet, the idea that often goes alongside such a conclusion – that Russia’s ability and desire to promote conflict centered on AI – is overly broad and possibly dangerous. Such thinking is reflective of an oversimplification common among non-specialists asked to incorporate AI into their view of future economics, politics and security dynamics.

In too many areas of policymaking and public discourse, AI’s expected impact on these things is reduced to a paradigm of “bigger, faster, smarter.” In such a view, hamstringing Russia’s AI fundamentals equates to diminishing capacity to confront the West. This overlooks the subversive character of international engagement on AI, which suggests that regressive conditions may simply push Russia toward new levels of adversarial behavior, particularly in cyberspace.

At the base level, AI instrumentation of cyber operations capabilities seems likely to be both more economical and accessible in the long run for a state strapped for resources. Developing AI applications means major data and computation costs, as well substantial investments in the training regimes, software products and more to translate strategic intent into tactical potential.

In cyberspace, Putin’s regime can continue to tap into private-sector developments of AI capacity for cyber defense and offense along several lines. On one hand, the traditional centrality of the Russian government in organized cybercrime has been jolted by recent events, but not fundamentally changed. On the other, powerhouse cybersecurity companies like Kaspersky will continue to be a fertile ground from which cyber operational possibilities might arise, particularly if such firms turn to partnerships with Chinese peers in response to constraining sanctions.

Russia can disrupt other AI efforts

More significantly, the role of cyberspace as a central avenue through which the training and functioning of AI tech operates will dictate an inevitable uptick in global cyber conflict. Much has been written about opportunities for poisoning AI systems or performing input attacks on the data used to train those same systems. Some such action can be taken outside of cyberspace. Researchers and security practitioners have notably used mirrors, stickers and other physical objects to trick AI systems in vehicles and sensors and alter their response to environmental stimuli.

The opportunities for manipulating sources of AI system inputs via the internet are immense, particularly given the surprising volumes of data used to train machine learning algorithms that are publicly accessible on repository sites like GitHub. As such, researchers have already forecast an uptick in cyber conflict activity beyond current trends to account for new interest in targeting AI systems. With Russia now falling behind competitors in the AI competition, such projection are more likely to be on point. After all, keeping up with the competition is often a task of degrading an opponent as much as it is of surpassing them.

This point belies yet another reason for believing that Russia’s AI setbacks will mean more assertive cyber aggression. Interference is the traditional bailiwick of Russian security services, and the strategic ambiguity promised by broad-scoped meddling with AI systems makes such tactics more appealing. What’s more, Russian interest in digital espionage, economic warfare, and political interference is likely to reflect a broad truth about the forthcoming AI revolution: As artificial intelligence rewires how societies function, so too will it determine the attack surface thereof. For a revanchist Russia, the value brought to Western economies by AI will likely be too tempting a set of targets to forsake.