EU regulation of Artificial intelligence in the shadow of global interdependence


China is one of the two globally dominant AI powers. That makes Sino-European regulatory cooperation an obvious project to pursue. At the same time, AI's military implications combined with geopolitical tensions, as well as differing visions on the role of technology in state-society relations cast their shadow. Project III asks: how and when should the EU collaborate with China in AI rule design and enforcement?

China looms large in the global AI landscape. Since the 2010s, the country has invested heavily in AI research and development, and it has alarmed many Western observers with its express ambition to become the global AI leader by 2030.

This Chinese presence impinges on EU AI governance in several ways: it raises competitiveness concerns, not least as a widespread AI application my boost Chinese manufacturing prowess and efficiency. In contrast, to US firms, leading Chinese AI firms (such as Baidu) have no strong presence in EU markets, and given political sensitivities surrounding European dependence on Chinese technology, it is unlikely that they will in the foreseeable future. 

That said, we should expect an increasing array of AI-powered devises to be exported from - not least products employing some form of computer vision connected to individuals, where Chinese companies benefit from relatively easy access to personal data, including biometrics. Some of these products may be innocuous enough in the application - think of medical diagnostics - but may have been trained on data that by European standards may have been gathered without respect for citizen rights. Import of such products into the EU would thus require some form of monitoring and verification regime, more or less coordinated with Chinese authorities. EU-Chinese regulatory cooperation will be inevitable as long as Chinese AI-driven products enter the EU market.

Many European politicians have cited extensive Chinese use of AI to monitor and control citizen behaviour, not least among the Uyghur in the Xinjiang province, as textbook examples of how AI can be abused to the detriment of citizens. 

While clearly cause for concern, such abuses do not in and of themselves challenge the effectiveness of EU rules directly. If anything, their effect is to strengthen the hand of AI-sceptical EU politicians. At the same time, they feed a wider worry that Chinese export of citizen-controlling technology around the world might entrench authoritative governments inimical to "European values" and less receptive to EU political entreaties. 

The appropriate European regulatory response, if any, is far from obvious. Would it entail attempts to convince Chinese authorities of export restrictions? Attempts to build a multilateral regime to restrict proliferation of surveillance tech?

At the same time, also in Europe worries about an AI-savvy China grow in light of rising geopolitical tensions. Europe's own military commitments in East and Southeast Asia are limited. So worrying as conflict there would be - with a Chinese invasion of Taiwan as the most obvious scenario - Europe would not immediately be involved militarily. Matters are different for the USA, and the prospect of military confrontation between the world two major powers remains scary, both for its direct and indirect effects.

Given most EU countries' ties to the USA through NATO, they are naturally wary of technological developments that endanger stability in the region. How, then, should European policymakers translate growing Chinese clout in AI into active EU-Chinese AI diplomacy, or standard setting efforts at other levels, for example in the context of the G20 or the OECD?

go to project IIIIIIIVV