02 Jan How the GDPR Will Help Europe Win the AI Race
In this blog, Benjamin Falk, co-founder of Yo-Da, outlines the current market landscape for personal data and provides insights on how data protection law will drive enormous socioeconomic benefits, bringing Europe in the lead of the international race for artificial intelligence supremacy.
I attended the GovTech Summit in Paris on 12 November 2018 and listened to Canadian Prime Minister Justin Trudeau opine that his country’s preference for strong privacy laws puts them at a disadvantage in the national race for artificial intelligence supremacy. Authoritarian countries like China, he argued, generate enormous data sets, because ALL the personal data of 1.3 billion people is accessible through a combination of Orwellian tactics.
This is an argument I have heard many times before and drives the global conversation around “AI Nationalism” to focus on a two-horse race between the US, whose Silicon Valley human capital advantage seems nearly insurmountable, and China thanks to its peculiar combination of size coupled with a complete lack of any privacy whatsoever.
But this logic is deeply flawed, and in fact, the exact opposite is likely to be true; countries that better protect consumer privacy will end up winning the AI race, and thanks to the EU’s General Data Protection Regulation (GDPR), Europe looks set to ultimately take the lead.
How exactly will this happen? By creating a more efficient and transparent market for personal data. AI’s are only as good as the data that feeds their algorithms. Put simply, the usefulness of the underlying data sets is of utmost importance in this race and, thanks to the GDPR, the depth, breadth, and quality of Europe’s data marketplace will be unrivalled.
By now we’ve all heard the cliché that data is the new oil. During the industrial era, as demand for energy exploded, a vast global market developed which helped ensure a reasonably efficient allocation of this scarce resource. But if data is the new oil, where is the marketplace? How do we know we are allocating it efficiently? Indeed, it is almost certain that we are not because we have no transparent price mechanism for personal data; we don’t allow people to price their own privacy.
Communism failed because the State fixed prices for everything, leading to resource misallocation. The same circumstance applies to China where the State is fixing the price of 1.3 billion of people’s privacy at zero.
This will undoubtedly cause tremendous distortions in China’s market for personal data. It’s human nature that our behaviour changes when we are being watched. Chinese people’s behaviour will reflect not just their true intentions, but their expectations of what is perceived to be positive. They will strategically obfuscate, opting out of data collection wherever possible, allowing only good signals to be collected while withholding negative signals. And as a result, the data sets in China will be severely biased. They won’t accurately reflect human decision making, and AI’s trained on these local data sets won’t be exportable outside China’s unique institutional setting. All the tech and cash in the world won’t make up for this fundamental constraint, indicating China won’t be the winner in this AI race.
The same logic applies, although to a lesser degree, to the US. But instead of the State setting the price of everyone’s privacy, we outsource that pricing decision to a handful of monopolistic technology giants including Facebook and Google. Every customer on these platforms is effectively forced to price their privacy at the same level, no matter what their true underlying preferences might be. The market for personal data in the US will be similarly distorted.
But Europe and Canada are leading the charge in a different direction by giving consumers rights akin to intellectual property control over their own personal data. You can think of the GDPR like copyright over any data that references you. And this is important because personal data is an information good, identical to a book, some software, or music. And if I am an author, and I sell you a copy of my book, you don’t own the copyright, you just own a copy of the book. You can read it, share it with your friends, and even write up a summary and post it online. But you can’t make money from my book or redistribute it. Authors receive royalties every time their content is reproduced. Intellectual property law is what enables the creation of a marketplace for information goods.
In the same vein, in Europe at least, data protection law enables the creation of a safe, transparent, consented marketplace for personal data information goods referencing 500 million people. Data sets that otherwise wouldn’t exist will be developed, and those observations will accurately reflect the behaviours of the people it references because this data will be exchanged voluntarily, and for mutual benefit. Consumers can now be paid royalties for their personal data information goods at a level concomitant with their privacy preferences, and as a result, MORE people will feel comfortable opting into this ecosystem. AI companies in Europe will have an advantage in terms of data quality and scope, helping to stimulate tremendous innovation in the European AI start-up scene. Counterintuitively, protecting privacy and giving people full ownership over all their personal data will drive enormous socio-economic benefits by encouraging an increasing volume of data sharing, not less. And we will have the GDPR to thank.
Benjamin Falk is a co-founder of Yo-Da.