Long Read

AI and the Disruption of World Politics

It is increasingly clear that artificial intelligence (AI) will be a foundation stone of commercial and state power in the 21^st century. A rush to develop AI talent and expertise, and a jostling for position is underway. The competition encompasses actors of all shapes and sizes and its outcome is difficult to predict. What can be said with a degree of confidence is that the competition itself, and the pattern of winners and losers it generates, will have a significant impact on the future shape and character of global order.

The idea that the major powers are engaged in technological competition is not, of course, new. What is new is the way in which the perceived requirements for successful AI development are interacting with pre-existing sources of state influence to create a different international dynamic. The US, for example, is concerned not only with growing Chinese technological prowess per se, but with China’s developing capabilities in AI in particular. This is both because it believes AI will be central to future economic and military power, and because it believes China has an unique combination of advantages when it comes to developing and applying the technology.

These advantages include the massive investments China has made in AI education and training in recent years, at every level of its education system. The mountain of data that its enormous population generates on a daily basis is creating a ‘Saudi Arabia of data’, the raw material on which AI algorithms depend for successful training. Add to that, the widely held view is that China has greater freedom to develop and implement AI applications unencumbered by Western concerns over transparency, accountability, personal privacy and surveillance.

Politically, concerns about China are now being layered onto a critique of its behaviour that existed prior to President Trump’s election, but have come to be more widely shared in the US foreign policy and political establishment since. The belief is that China’s rise has been built on unfair and often illegal practices, including intellectual property theft and forced technology transfer in return for US company access to the Chinese market. Protectionist measures designed to enable import substitution figures on the wider charge sheet.

So while the US may still be ahead in AI development overall, it is now suspicious of China and fearful of its potential. These fears are actively being turned into measures to combat China’s rise. For its part, China resists these claims about its previous behaviour, seeing concerns over its current technological trajectory as a disrespectful US attempt to keep the country down while preserving Washington’s own position of global leadership.

This US-China bi-lateral dynamic is, however, only one aspect of the wider jostling for international power and position that the AI revolution has unleashed.

President Putin, aware that Russia is behind in the AI race, has recently sought to deepen Russia’s technological cooperation with China. He has allowed Huawei, the company at the centre of the US-China technology row, to massively increase its footprint in Russia. Having just introduced a capacity to cut Russia off from the global internet, he is also interested in learning from the Chinese how best to use AI for domestic surveillance and social monitoring purposes. The Russia-China relationship is a complicated one, but both Putin and Xi Jinping clearly see technology cooperation as an opportunity to advance their own agendas while reducing US technology leadership and global influence.

The EU has also seen a possible path to global leadership in AI and is pursuing it. It is currently behind the US and China in terms of investment volumes, number of AI companies and ability to translate research into real applications. Companies and researchers in its member states are also at a disadvantage in terms of the amount of data they can access in domestic markets.

The EU believes, however, that it can use its good research capability in AI, the size of the single European market, and a unique emphasis on ‘trustworthy AI’ to shape the future. Chancellor Merkel hinted at this in comments she made in 2018: “In the US”, she said, “control over personal data is privatised to a large extent. In China, the opposite is true: the state has mounted a takeover.”

The argument here is that the EU should seek to occupy the space between the two by offering a well-regulated model of AI development and adoption focused on privacy, transparency and accountability that can win public trust. On that basis, the EU’s play will be to attract ethical investment and talent, as the foundation stone for it becoming a competitive global leader and standard setter in the new technology. As such, it is pursuing an approach that exemplifies its wider preference for lawfare not warfare when it comes to trying to leverage power and influence on the world stage.

The competition among major powers over AI raises difficult questions for other states too. Middle powers like the UK, France, South Korea, Japan and Australia are already facing difficult choices with regard to which superpower partner to choose when it comes to AI collaboration. Smaller states risk being side-lined completely by a technology battle in which they simply do not have the resources to compete.

Disruption Ahead

This landscape portends significant disruption.

The US, Chinese and other global militaries are actively exploring what all this might mean for the future of warfare, the concern being that the state that emerges as leader in AI will emerge as the world’s pre-eminent military power. This raises the significant prospect of a real arms race, huge difficulties for any attempts at arms control, and question marks over the ability of the major powers to effectively manage periods of crisis and high tension between them, especially in a world where AI facilitated autonomous weapons may be deployed at scale.

A new world of export controls on sensitive and dual use technologies is emerging. We can expect new rules on inward investment in the technology sector and overt pressure from major powers on allies not to adopt technology developed by an adversary. This will have significant disruptive effects on markets, global supply chains, corporate licenses to operate and innovative cross-border research partnerships. The impact on higher education could also be profound, given the importance of the human capital at the heart of AI development, and increasing concern around training the citizens of a potential adversary.

There is significant potential for the world to be divided into new technology driven spheres of influence. One gets a clear sense of this in relation to China’s use of surveillance technology. China not only uses AI to perfect the surveillance and social control of its own population but is engaging in the export of those tools to other regimes with an authoritarian bent around the world. Venezuela and Ethiopia are just two examples. This adds an ideological dimension to the raw competition for economic and military advantage ongoing between the US and China, raising profound questions about what a Chinese dominated world might look like. Reinforcing this tendency for the world to divide into competing blocs, there will be a major power battle to draw partners into US, Chinese and EU led technology eco-systems and regulatory norms for the long-term.

The outcome of this will be difficult to read, because it will also depend on the choices of actors too often overlooked. While the middle powers will be squeezed in this environment, for example, it would be a mistake to see them as mere bystanders. The decisions these countries make with regard to which AI superpower to align with will have a big impact on the future global balance of power. Old alliances still matter in determining outcomes, but President Trump’s erratic behaviour also makes relationships more fluid than they otherwise might have been.

More than that, all of these states are making their own investments in AI to solve specific problems, develop specific capabilities and hedge against AI enabled economic and military breakthroughs by others. Some of them may emerge as significant centres of AI expertise and influence in their own right. Collectively, they will be influential in helping to shape what rules and norms, if any, emerge across the world with regard to AI development and use. It will be their decisions that help give such norms and rules legitimacy and reach.

Even small states may be able to find significant niche roles and positions of influence in the emergent order. Estonia, one of the most connected countries in the world and a hot bed for tech start-up activity per head of population, has been addressing legal issues at the forefront of the AI revolution, removing or effectively managing points of tension and friction well ahead of others. It may set precedent for others to follow and is emerging as a leading centre for the practical application of AI as a result.

At the heart of the AI revolution are corporations such as Apple, Google, Alibaba and Tencent. They have massive resources that outstrip most states, spending tens of billions of dollars on research and development every year. They have access to pools of human talent and data that most states can only dream of. They will come under major pressure to align with their ‘home’ governments in a fragmenting world but have the capability to be political actors in their own right and to slip national moorings in pursuit of their own corporate interests. The products they choose to develop can bring huge benefits to humanity but may also unleash forces that destroy social trust, reduce freedom and are difficult to control. Their decisions on what and where to invest, will be pivotal not only in shaping what kind of AI emerges in future, but also to the inter-state balance of power.

If anyone does not believe terrorists will find a way to use AI, then they should think again. Terrorist groups have a long track record of turning new technologies to their advantage. They have become adept at using social media platforms to reach and train new recruits and to plan attacks, often behind the cloak of sophisticated encryption tools. In recent years, they have shown a willingness and capability to use small commercial drones to send bombs to targets in Iraq and elsewhere.

This pattern of technological adoption and adaptation will surely continue in the AI era. It would be a failure of imagination to believe terrorists will not use social network mapping tools to better understand and target groups of connected officials and influencers. As the price of small AI-powered drones comes down to what some suggest will be the price of a smart-phone, their use to carry out targeted assassinations will likely grow. Autonomous vehicles will be used to deliver car bombs to target destinations. Chatbots will be used to automate the recruitment process, allowing it to operate at scale. Deep fake videos will be used in the manufacture and dissemination of terrorist disinformation and propaganda.

In this world of AI led disruption and jostling for power, attempts to cooperate internationally will be key if the worst outcomes are to be avoided. But while there have been some successful international attempts to agree the principles that should underpin AI development and use, principles are a far cry from effective international regulation. National regulators are doing a better job in some locations but obviously have limited reach.

In terms of international politics, AI is an unregulated field. Navigating that reality, exploring what might come next and what the extreme consequences may be is a challenge for all.

November 2019