It's rare that I agree with Vladimir Putin. But he was spot on when he said whoever cracks artificial intelligence first will run the world.
The assumption in Washington, and among its allies, is that Silicon Valley will win the AI race for Uncle Sam and, subsequently, for the West. That attitude is dangerously complacent and may not end with "American First." Silicon Valley's big tech firms are heavily investing in AI, but collaboration between public and private sectors in the U.S. is lacking, and investment and talent is skewed toward big tech. China is catching up with America fast, and it could well emerge dominant.
Get what matters in tech, in your inbox every morning. Sign up for Source Code.
Meanwhile, Europe currently lags the two superpowers of America and China by a huge margin. It lacks China's access to data, low scruples, and the state firepower of centralized market control that are helping the People's Republic to make huge leaps in using the technology. And it lacks access to funding when compared with venture-capital rich North America: R&D spending in Europe was €2.4 billion to €3.2 billion in 2016, compared with €12.1 billion to €18.6 billion in North America.
This week, the European Commission took a first shot at rebalancing the scales. President Ursula von der Leyen announced a series of policy proposals aimed at increasing data access in Europe, laying the groundwork for human-centric AI, and advancing the so-called digital single market. For the moment, it's unclear whether these initiatives will take on a protectionist nature or whether they will create bridges to give Europe the boost it needs. Different elements of the strategy point in both directions.
These proposals are part of its attempt to forge a third way on technology that helps it regain its "digital sovereignty" in light of increasing distrust of its allies and its enemies. We see this wariness in the ongoing debate about engaging Huawei to build 5G infrastructure, as well as Europe's handling of the Silicon Valley giants' feigned pitch to be regulated. As a consequence, Europe risks missing the bigger picture.
Europe's nascent digital sovereignty is not best served by opening a third front in the battle. Instead, the EU should play to its strengths and build a Western partnership model across public, private, research and academic sectors for the world's democracies to win the AI race together.
As was the case with data flows and data protection, the EU is placing an emphasis on the ethical and human centric aspects of AI. China makes no such distinction. Beijing sees the national security potential: from using AI facial recognition to track minorities to developing autonomous weapons that select targets with no human input. It sees AI's potential free from the legal and ethical constraints of the West. As a consequence, a new silicon curtain between democracies and autocracies has been created. Allowing China to develop an AI edge could greatly threaten our security, privacy and freedoms for years to come.
Where should we start in preventing such a scenario? The U.S. has formed a National Security Commission on AI; the Western Alliance as a whole should participate in such exercises — first NATO allies, and later partners like Japan, and other like-minded nations in the Pacific, Asia, Africa and South America, to make recommendations. This broader alliance of like-minded democracies should find a formula for the exchange of data, potentially building on the EU's model of data adequacy agreements, to enable freer flow of data between allies.
Europe should also significantly increase its investment, including in NATO. We regularly hear of NATO's 2% target — where allies committed to spending 2% of GDP on defense. But we have also set another target of 20%, the amount that should be spent on new equipment and R&D. Let's increase it to 30, and let's see President Trump tweeting about this target, too.
Entering into such a partnership model should not mean that the EU blindly accepts a tech sector where the FAANGs — Facebook, Apple, Amazon, Netflix and Google — absorb or eliminate all competition and we allow data to be processed in an unscrupulous manner. The EU and the U.S. would have to learn how to do business with one another where data exchange is concerned. We are long overdue a comprehensive and all-encompassing trans-Atlantic data-sharing agreement that both sides are enthusiastic about.
Western democracies should see this as an opportunity to set down the standards for establishing global conventions that will regulate the development and use of AI for years to come. Such a move would reflect the wider partnership model adopted by Europe over the past 70 years — a model that has only ever made us stronger.
In the field of international relations, we understand that multipolar worlds are more unpredictable and prone to conflict. A Technological Alliance of Democracies — one that ushers in a more stable bipolar system — could win the race for AI while preventing the inevitable race to the bottom in the pursuit of marketplace dominance.
Despite the many political differences among Western democracies, we need only ask ourselves one question: Would we prefer a Communist dictatorship to win this race? Rather than pursuing our own individual vision of a perfect and sovereign AI model, we should instead work to establish an imperfect system that actually becomes a reality, and which serves the purposes of our democracies, our people, and the pursuit of peace.