Is Europe going to declare war on (artificial) intelligence?
By Alexis Normand on Nov 27, 2017
[Ed. Note: The following column originally appeared in Le Monde. It has been translated from French.]
Will the General Data Protection Regulation (GDPR), which will come into force in early 2018, sound the death knell for our innovative industries? Here’s a quick analysis from the world of GAFA (Google, Amazon, Facebook and Apple).
Like many French people living in the US, I am amazed to find that we can’t avoid even the most private aspects of our lives being tracked. To buy a car, rent a house, even take out insurance, interest rates can vary greatly depending on a credit history that’s meticulously compiled. Credit scores don’t limit spending so much as encourage lending. Max Weber pointed out this relationship between Protestantism, transparency and the spirit of capitalism – a good Protestant has nothing to hide. For fans of smart objects, like myself, self-monitoring is a completely personal choice. My smart scales warn me of sudden increases in arterial stiffness. My Alexa speaker takes note of the music I like. My smart vacuum cleaner maps out my home. All of this monitoring, often seen as an obligation, is the condition for new advances.
The assumed monetization of transparency is the big driver behind the digital revolution. Eric Schmidt, the CEO of Google, said it best while defending the intrusive nature of his company's services: “If you have something that you don’t want anyone to know, maybe you shouldn’t be doing it in the first place.” Transparency creates trust which is conducive to exchanges and opens up new marketplaces. For example, I’ll only rent out my home on Airbnb to recommended guests, and I only take an Uber with the best-rated drivers.
The difference in American and European leanings is creating an economic stand-off. On one side of the pond, the Old Continent is turning out to be unable to produce digital champions, with our leaders trying to tax GAFA late in the day due to not knowing how to monetize data at its source. As for the US, they seem unable to avoid breaching trust. In the latest of a series of scandals, Equifax, a major credit rating agency, had its systems and the data of 145 million Americans hacked. Although aware of the problem in July, the company only acknowledged it in October. Worse still, Facebook realized that Russian hackers had been able to provide extensive funding for targeted advertisements supporting the Trump campaign. Surprisingly, these scandals resulted in little legal action in the US. No one questions the supremacy of GAFA in their native country.
Conversely, Europe is taking the bull by the horns. The General Data Protection Regulation (GDPR), which will come into effect in 2018, is forcing platforms to adhere to new obligations. To avoid abusive data collection, the regulation sets out that companies must obtain “clear and explicit” consent and imposes a principle of proportionate use (i.e. data used is limited to what is necessary, nothing more), thus obliging platforms to collect the minimum amount of data. The GDPR also entails a “right to be forgotten”, meaning an individual can ask for personal data to be deleted. Companies will have to provide a full record of data processing, a procedure that comes with a considerable extra cost for start-ups that share data. There are sizeable, and very off-putting, sanctions which can amount to up to 4% of a company’s turnover, a significant increase on the CNIL’s maximum fine of €150,000.
In the US, this new framework is seen as an insult to intelligence, artificial intelligence that is. Obtaining consent represents a loss for databases used for iterative learning in so-called machine learning solutions. These software programs are different in that they can make decisions. However, the GDPR gives citizens control over “algorithm profiling”, in other words, any decision made based on a profile. Elon Musk highlighted the legal uncertainty weighing on the most ambitious developments. For example, driverless cars will decide for us how best to avoid an accident. In the world of connected healthcare, Carmat has developed an artificial heart that maintains vital functions by automatically rebalancing parameters. How far will a machine have to justify its “reasoning”?
It would seem the battle has only just begun, and progress in artificial intelligence is increasing the stakes. Thanks to expertise acquired through data, Google has just announced the launch of headphones that can instantly translate 40 languages. This spectacular technology can be applied to healthcare. In the US, the Precision Medicine Initiative, to which President Obama granted $215M in 2015, relies on unlimited digitization. From analyzing genes to monitoring vital signs using sensors, the Initiative posits that data analysis will help stratify patients and pinpoint therapies. Meanwhile, silos imposed in Europe, especially on DNA sequencing, risk causing the continent to miss out on this revolution.
Can Europe catch up without abandoning its principles? Protecting freedoms is not to be sacrificed, not even to keep up with the competition. So efforts must be doubled. Unlike in the US, centralizing systems, especially in healthcare, can clear the way if public stakeholders think like platforms. Similarly, some new rights may be an opportunity for fairer competition. Portability, for example, makes it possible to download and transfer data from one platform to another, like changing your phone carrier. This is lacking in the US. All that’s missing in Europe is a regulatory framework. The US does, though, have the capacity to match any new constraint with massive support for investment.