Friend or Foe? Here's how the EU will regulate AI

I have a confession to make. I don’t really like Artificial Intelligence.

Or rather: I don’t really like the term “Artificial Intelligence”.

First, because there isn’t one AI but a multitude of applications for Artificial Intelligence.

Second, the term 'Artificial Intelligence' has too often served as a receptacle for certain popular beliefs or fears in our society, largely fueled by the cinema. The little anthropomorphic robot Wall-E has not succeeded in making us forget the T-800 from Terminator... But we must not stigmatise technology.

Artificial intelligence is the combination of three elements, over which we have complete control: data, algorithms and computing power.

These are the points on which we can act - the quality of the data, the transparency of the algorithms in particular.

Our guiding principle is always to encourage innovation and applications for the benefit of our citizens, our start-ups and our industry.

In February 2020, the AI White Paper launched a wide public consultation which received over 1200 contributions: the interest in the uptake of AI is real and the potential of AI applications is immense in fields as diverse as: health, transport, energy, agriculture, fashion, tourism and (cyber)security.

However, like any technology, AI can also lead to risks and abuses, such as social scoring or possible discrimination in a recruitment process.

This is why our approach on AI is as concrete as possible: we must invest in European excellence, ensure that AI in Europe is developed according to our rules and standards and finally, exploit the full potential of AI for industrial purposes. 

Invest in European excellence

In terms of investment, the Commission will invest from its Digital Europe programme and Horizon Europe around €1 billion per year.

This European public investment will be able to generate complementary private investment that will allow us to propel Europe to the forefront of technology.

Our goal is to reach a collective annual investment volume per EU and Member States of €20 billion per year over this decade, also thanks to part of the approximately €140 billion that will finance NextGenerationEU's digital investments.  

Ensure a predictable legal framework

Europe has clear rules in place to provide legal certainty and consumer protection when it comes to goods and services. Today, we are the first jurisdiction in the world to propose a comprehensive legal framework to do the same for AI.

Behind all the complexities, our proposal is based on a simple approach: ban the unacceptable, impose strict conditions when the risk is high, and ensure transparency for AI systems which represent minimal risk (the vast majority).

We follow the proportionate precautionary principle and will ensure that AI in Europe respects our rules and values. There, high quality data sets and testing play a central role.

We want AI systems placed on our market to be accurate, reliable, safe and non-discriminatory – regardless of their origin. In order to ensure, for example, that a robot-assisted surgery is safe or that software sorting CVs in a recruitment process does not discriminate, we need high quality and representative training testing datasets. That is why we have included the high quality of datasets among the key requirements to put high-risk AI systems on the market.

Companies will benefit from our approach too. Last year, about two thirds of EU companies saw a lack of trust among citizens in AI. By promoting trust in AI, we create the conditions necessary for companies to start investing.  

We are also updating our rules on machinery products, by enhancing even further the safety of users and consumers of machines and robots, while making sure industry – including SMEs – can thrive and innovation flourish.  

Make the most of Artificial Intelligence for our industry

A few weeks ago, we presented the Digital Compass, where we set the ambition of 75% of European companies to be using data, the cloud and AI by 2030.  

Very soon, we will present our update to the March 2020 Industrial Strategy. All of this is interrelated.  

To remain ahead of the curve in Europe, we need to develop suitable European infrastructures allowing the storage, the use, and the creation of data-based applications and Artificial Intelligence services.  

This is the reason why today’s proposals should be read together with our actions to increase investment in data and cloud, and in computing infrastructure. 

The right investments and framework conditions for key technologies will shape our industry of the future, and therefore the future of our industry, and of our society as a whole.