How sanctions and security-first policies hamper global cooperation

Geopolitical multipolarity is, in many ways, the democratization or decentralization of hegemony. While it may be argued to be morally laudable, the shift we see also carries profound risks across regions and sectors.

Policymakers and businesses are having trouble finding footing on a political terrain that shifts with every step. Recently, secular forces of deglobalization have tore at the fabric of the global economy, leaving states and companies exposed to these abrupt changes.

As the new global paradigm of security-first frameworks is implemented, the cascading effects will hamper multilateral cooperation, especially in areas where global cooperation is fundamental, such as a coordinated energy transition and AI governance.

Embedded in a security-first paradigm are a set of policy tools that are inflationary by nature. These tools primarily include sanctions, secondary sanctions and export controls.

Over the past 20 years, the United States has imposed more sanctions than the United Nations, European Union and Canada combined. This trend is not slowing down; in fact, it’s accelerating.

Between 2017 and 2020, US President Donald Trump placed more than 3,900 people and companies under sanctions. This starkly contrasts US President George W. Bush, who imposed 3,484 sanctions on entities and people in eight years. But this is not unique to party lines. US President Joe Biden has also imposed a series of export controls and sanctions during his tenure – and he could still sign more.

As author and geoeconomics expert Agathe Demaris writes: “Sanctions are a low-cost policy with high political dividends: public officials who impose them tend to boost their approval ratings. Furthermore, the burden of sanctions does not fall on policymakers but on multinationals and banks who have to adhere to them.”

These measures are implemented under a security-first policy framework for nation-states seeking to enforce sovereign safeguards in a new political climate of elevated hostility and distrust. Inflationary spikes across regions and sectors are, therefore, a natural byproduct and the internationalization of geopolitical tension means it has global ripple effects.

Balkanized power means standardization; cross-border protocols for trade and data become fragmented as each state competes to implement and maintain its model first.

However, despite these risks, export controls and primary and secondary sanctions will entrench themselves as the modus operandi of policymakers. This is notwithstanding compelling evidence that historically, sanctions have not been very effective and in some instances, have done more harm than good.

A comprehensive analysis of US sanctions programmes from 1970 onwards reveals that only 13% of targeted nations adjusted their actions in the manner the United States intended.

Balancing legitimate national security concerns on a national basis cannot be disregarded, though a more flexible framework may be needed to achieve global objectives. Two themes of great consequence are the energy transition and the governance of AI. Both require global cooperation, for a lack of it will almost certainly result in macroeconomic or political insecurity.

At the heart of these two themes is the US-China rivalry. As the world’s two largest economies, bilateral relations do not remain a narrow dichotomous dynamic but ripple out globally. China plays a key role in the global supply chain for critical minerals, specifically rare-earth resources. The Asian giant is also a major contributor to the energy transition.

China dominates the market for electric vehicles (EVs) – in 2023, the Asian giant accounted for approximately 60% of total global EV sales, with approximately 11.5 million units sold annually. In comparison, the United States sold about 1.4 million units, a more than 40% increase from the previous year but still only about 10.2% of global EV sales.

Equally for solar panels, the sun shines bright on Beijing. In 2023, China installed 170 gigawatts (GW) of new solar capacity, which accounted for around 43% of the global total of approximately 400 GW. The United States, on the other hand, added about 20 GW of new solar capacity in the same year, accounting for about 5% of global solar installations.

US officials have cited security risks with Chinese EVs and until those are investigated and resolved, integrating them into the US economy is not likely. Both countries are deploying targeted industrial policies to make their EV products more competitive globally. Asymmetric state subsidies and support are also being incorporated into the economic security model, leaving policymakers at a standstill.

But equally as important – if not more so – is the global regulation of AI governance. While there may be a difference in domestic deployment with varying regulatory parameters, there are areas where international consensus can – and should – be reached. Primarily, this applies in the military domain when the fog of war or the prospect of it should not be delegated to AI.

Like the Geneva Conventions, a global framework for using AI in military contexts would likely help states avoid falling into a type of "Thucydides Trap," yet, in the context of technological advancement and AI – what I call a “Promethean Trap.”

Specifically, countries with strategic rivalries engage in a technological arms race that is highly likely to lead to conflict. A parallel here can be drawn with the Cold War and the resulting proliferation of nuclear weapons.

Before relations began to deteriorate, China and the United States worked closer together on AI than any other group of countries. While security concerns must be considered and strategic information preserved, there are areas of cooperation that are ripe for collaboration – specifically, AI governance as it relates to war and its usage in nuclear scenarios.

War is not inevitable but if tensions continue to rise, technological guardrails must be established to avoid a terrible conflict – one in which humans are the primary controllers and not an algorithm.