'Please check against delivery'

Ladies and gentlemen,

The coronavirus has certainly disrupted a lot of plans this year. If 2020 had been an ordinary year, we would likely have been meeting today. It is nevertheless a pleasure to speak to you about artificial intelligence today, at the very end of the Croatian Presidency. The last few months have certainly been the most unusual, the most unsettling environment in which a Council Presidency ever had to operate. So I am glad that you found a way for this conference to go ahead.

And this is possible, of course, because of digitisation. The Covid crisis has challenged each and every one of us. Our health and health systems, our economies, our very personal way of living. It has also been a testbed for digital solutions. As we have had to avoid physical interaction, digital technology has stepped in to fill the gap. It has allowed us to go – and to shop – online. We have communicated online, worked from home, taught our children from the distance, and followed sports classes on our computers. And that experience is a powerful reminder of just how much digital technology has to offer us.

Digital technologies are also able to help us deal with the virus itself. Data, in combination with artificial intelligence and supercomputers, are major assets in detecting patterns in the spread of the virus or potential treatments. This is why the European Union has been funding a range of research and innovation actions in the field of AI and health.

Some three months ago, we also launched an initiative to collect ideas about deployable Artificial Intelligence (AI) and Robotics solutions that could help face the ongoing COVID-19 crisis. In response to this call, numerous digital solutions were shared on our website.

I am often asked if all the benefits that digital solutions have brought to us during the crisis have changed the Commission’s digital policy. Have we been too critical of tech companies, too careful and bureaucratic when it comes to digital technologies?

My answer to this is that our objectives are more relevant than ever. The more we use and depend on digital technologies, the more important it is that these technologies are in tune with our values, beliefs and rules.

We firmly believe in the great potential that the digital transformation holds for our societies and economies. At the very beginning of our mandate, this Commission identified what we call “the twin transition” as our main strategic objective. An objective to promote the transition to both a green and a digital economy and society.

This has proven more relevant than ever in our strategy for recovery after the corona crisis. Together with our proposals for a modern and revamped long-term EU budget, we have proposed a new recovery programme. We called it Next Generation EU.

I think the name for this programme is well-chosen. As we suggest borrowing money from our children and grandchildren, we should invest in a way that benefits them. To make our planet greener, and to invest in modern technologies that will secure their jobs and prosperity.

And there is huge economic potential in investing in our digital infrastructures and capabilities. We need first-class connectivity, efficient cloud solutions and computing power in Europe. We need to make sure that the vast amount of data that is created every day – many of which are non-personal data – are made available to innovators and start-ups. And we need to invest in research and skills.

There is even more to the story. When we adopted our digital strategy, back in February, we said we want to “shape” Europe’s digital future. We want technology to serve a true purpose, to make a difference in people’s lives. We want digital technologies to work for each and every one of us.

Artificial intelligence is a case in point. If properly developed and used, it can work miracles, both for our economy and for our society. It can help us better manage health crises like the current one – and predict the next one before any human being can. It can help us to identify fraudulent transactions on e-commerce market places, so that we do not end up buying counterfeit products without knowing it. And it can help us become climate neutral by 2050, by saving energy and waste at any stage of the product cycle. We need to put all our joint effort into developing these European AI solutions.

But artificial intelligence can also do harm. It can lead to discrimination, amplifying the prejudices and biases that are already inherent in our society. Women might not be offered a top management job because they have occupied fewer such jobs in the past than men. Immigrants and people belonging to certain ethnical groups might be targeted by predictive policing techniques that direct all the attention of law enforcement to them. This is not acceptable.

Artificial intelligence may also alter the behaviour of products after we have bought them. If this means that a music streaming service to which I have subscribed recommends a song to me that I don’t like, I can live with this. But if my self-driving car takes decisions on whose life to spare in case of an accident, I would want greater scrutiny over such decisions.

This is why we said, in our White Paper on AI, that high-risk applications need to be properly regulated. That we need to know they are technically robust and safe. That they have been trained on reliable data. That, somewhere, there is a human in the loop.

I am aware, of course, that it is not easy to define which AI systems are high risk, and which are not. For some areas it may be obvious – for example, when AI might be used for mass surveillance purposes. For other areas, this is more difficult.

This question was also one of the key questions raised in our public consultation on the White Paper, which ended in mid-June. Close to 1200 people or organisations responded to this consultation – which shows how much people care about this issue.

Most of these contributors agreed that AI, if not properly framed, might compromise our fundamental rights or safety. Many of them agreed with us that we should focus our attention on high-risk applications. But rather few of them were convinced that we had already found the silver bullet that allows us to distinguish between high and low-risk applications.

This is why today’s conference is also important. Given that AI, one way or the other, affects every one of us, we need a thorough debate on how to deal with this technology. We want to develop European AI with clear rules and innovative solutions to boost our economic growth and societal welfare. For this, we need to safeguard our trust in this technology. Because, without trust, no sustainable progress is possible.

In this spirit, I wish you an inspiring and productive conference!