Australia and the European Union need to collaborate to promote the adoption of inclusive, safe and sustainable AI in the South Pacific.
While the United States and China compete for global AI leadership, the EU has opened an alternative rights-based route in the ongoing battle of AI norms, which seems to resonate with Australia’s vision and interests.
These conversations are making their way to Pacific island countries that are starting their AI adoption journeys. The technology could help address some of their most pressing challenges, but as a region at the crossroads of geopolitical influences they will have to take their pick in the battle of AI norms.
Pacific island countries are starting to implement digital strategies and have ‘a late-mover advantage to selectively adopt best AI practices from the US, China and the EU’, according to a report by the AI Asia Pacific Institute. The Pacific is already a geostrategic battlefield of the China-US rivalry and AI emerges as the latest arena for influence.
AI will be at the top of local agendas as it offers Pacific countries unique opportunities to address non-traditional security challenges. It can aid the fight against illegal, unreported and unregulated fishing; support research on climate change; improve natural disaster management; and even preserve culture through revitalisation of indigenous languages.
Pacific island countries could also be affected by the lesser-known consequences of AI, particularly its environmental impact. Energy-ravenous AI systems drive up power demand, exacerbating climate change effects. The region is already on the frontline of global warming. AI adoption should be accompanied by a rapid transition to renewable energy sources.
Pacific island countries face a choice of existing AI models to draw from.
The US has shown the world its focus is now on retaining global competitive edge in AI at all costs, resorting to deregulation, tariff barriers, trade bans, and diplomatic pressure—although high-end Nvidia chips still make their way to China through intermediate destinations.
China, meanwhile, opts not only for an interventionist framework, but also for an insidiously pragmatic approach. China has chosen to develop smaller and more specialised language models. Using unconventional tactics including smuggling and espionage—for which a Google engineer faces charges in the US—Beijing relies on supposedly smaller budgets but expects strategic and targeted impact.
Between these two geopolitical and normative poles, the EU emerged last year as a creative force. In its search for strategic autonomy and to appear as a credible normative power, it adopted the AI Act after months of negotiations. This safety legislation introduces different obligations for AI products and services sold on the EU market, adjusted to their level of societal risks. To avoid stifling innovation, the EU paired the legislation with an initiative to mobilise €200 billion for AI development.
Pacific island countries could be tempted to draw inspiration from EU norms. Certainly, the process is ongoing in the French overseas territories. If the EU’s AI Act eventually becomes applicable there—which first requires a legal translation into domestic French law first, as happened with the EU’s Digital Services Act and General Data Protection Regulation—this could serve as a practical and peaceful bridgehead in the broader region.
Australia is also acting in this battle of AI norms in the Pacific. For example, the Department of Foreign Affairs and Trade organised a series of side events on AI adoption in the Asia-Pacific at the Paris AI Action Summit.
In this situation, and as raised last week by the President of the European Commission, it is necessary for Canberra to work more closely with Brussels, a partner with a similar vision of what AI governance should look like.
For Australia, defence-related AI is covered under the AUKUS agreement, but those partners diverge on broader governance approaches. For example, Australia signed up to the Paris AI Action Summit statement on inclusive and responsible AI, but the US and Britain did not. The EU did.
Moreover, Australia seems to be on the path to developing its very own AI legislation, modelled after the European risk-based approach while simultaneously focusing on interoperability and technical standardisation.
Pacific island countries could also decide to work towards a consensual regional approach to AI governance. Indigenous technologists in Australia and New Zealand are already making a push to include indigenous data sovereignty frameworks in global AI governance discussions. The next Pacific Islands Forum in September this year could be the platform for such discussions, as AI is recognised as a driving force for change in the 2050 Strategy for the Blue Pacific Continent.