Check against delivery.

Ladies and Gentlemen,

Do you remember the first episodes of Star Trek in the 1970s? I thought they were truly awesome – particularly because of the technological gadgets that Captain Kirk & co. used to communicate with each other or to kill time while travelling through outer space. These included a mobile communication device and a personal access display device – PADD for short – which crew members could use to watch videos, read diagnostic reports or send messages. Well, the digital world we live in today has certainly surpassed 70s science fiction in many areas.

One industry in the midst of a rapid digital transformation is the financial industry. This transformation is driven by many different elements, and artificial intelligence (AI) and distributed ledger technology (DLT) tools in particular are currently at the centre of people’s attention. Ever-growing data volumes and ever-better ways to make use of these tools are enabling companies to launch completely new products, services and business models. We are undergoing a period of radical change, prompting many new questions for both market participants and supervisors and regulators, too. And, of course, we need to develop answers to these questions.

After all, regulators and supervisors are tasked with nothing less than ensuring the proper functioning, stability and integrity of financial markets, also in times of digital transformation. Our continuing mission: to explore new digital worlds, to seek out new complex legal issues and challenges, to boldly go where no regulator or supervisor has gone before. In doing so, let me focus on four fundamental questions.

Who will be at the receiving end of supervision in the future?

The first question is: who will we be sending letters and e-mails to in the future? In comparison to work processes in other industries, a relatively high share of the processes underlying financial services still run internally. This means that a lot continues to be done in-house, and work is occasionally outsourced to third-party providers. However, outsourcing has gained momentum in the financial industry, too, and this development will pick up even more speed in the future. If we follow this train of thought to the very end, we cannot rule out the possibility that only a few tech groups will be providing services for the entire financial sector at some point.

Of course, this would be a different financial world to the one we have known so far. But as I said before: our duty to ensure the proper functioning and stability of financial markets will also need to be fulfilled under such conditions. This raises a question that we already need to think about now: does it make sense to set out specific and direct requirements for certain service providers, especially cloud service providers? One thing is already certain: purely national solutions won’t be enough in the long run. The financial industry is already connected to such a degree that we need European or, ideally, even global approaches. Fortunately, we have already got the ball rolling. The European Supervisory Authorities (ESAs) have advised the European Commission to develop an appropriate oversight framework for monitoring third-party providers that are critical for relevant entities.¹

Is the principle of human responsibility safeguarded?

The same question – who should we actually supervise? – will be reinforced by the growing use of BDAI applications. Financial institutions are increasingly transferring entire processes – and, in certain cases, decision-making – from humans to machines or algorithms. For the Star Trek aficionados among you, this is nothing new. In the series, Data, an android, was even a starship commander.² And if we look at the real world of finance, we can see that insurers, for instance, are already capable of carrying out processes without human intervention, e.g. when assessing risks in new business or processing claims. What started with standardised processes has now progressed intom more complex matters. It is evident that it will be possible to delegate many other processes to alforithms in the future.

But what happens if something goes wrong? Would it be acceptable if a company’s management told us: “It wasn’t me – it was the algorithm”? I believe the answer should be no! For this reason, we must do everything we can to uphold the principle that humans must bear responsibility, even when artificial intelligence is used.

The ultimate challenge that comes with the question of who we should supervise is posed to us in the context of open blockchain technologies. Who is the ultimate bearer of responsibility in a context which is by definition based on a more or less anonymised network? It goes without saying that this is particularly relevant with respect to anti-money laundering (AML) requirements. If such technologies are to prevail, both the industry and regulators will have to find answers, soon.

How much is the time frame for supervisory decisions changing?

Let me focus on another quite fundamental issue: How long will supervisory approvals remain valid for in times of artificial intelligence and machine learning? The licensing process is based on the ability to reliably identify what is in and what is out of the regulatory framework – and the time frame in which this process takes place is changing. Let us take internal models for determining regulatory capital requirements as an example. And let us assume that they are – at least in parts - driven by self-learning systems and that they are evolving continuously. In such cases, an internal model that has just been approved may look quite different barely after the ink on our approval has dried.

Regulators and supervisors need to be prepared for such a scenario, too. Ideally, this should be done by defining in advance how much a model can change without the bank, insurer or investment services enterprise needing to get straight back in touch with the supervisor. So far, supervisory model reviews have largely concentrated on model design and input factors. The question could arise as to whether we might have to turn to procedures that focus on the output or an output corridor – i.e. on the impact.

Will money be a private matter in the future?

Last but not least, I would like to invite you on a little journey to Ancient Greece following our voyage into outer space. In Athens, coining was for a long time in the hands of aristocrats who owned large swathes of land. Aristotle or one of his disciples wrote a report on the major coinage reforms introduced by Solon in 593 BC³. For some experts, this was a milestone in terms of the state’s privilege to create money in Europe.⁴ In Central Europe, it was Charlemagne who ordered that coins were to be made only by minters in his court⁵.

Statesmen from ancient times and German emperors are a thing of the past. But issuing money is still a prerogative of government institutions – or, to be more precise, it remains in the hands of central banks. It is in very good hands there and should stay there, as the value of a currency is still based on the trust that other people have in it. Only public authorities are able to maintain this trust over long periods of time.

Now of course, there are people who are suspicious of both government institutions and traditional intermediaries such as banks and central counterparties in investment trading. For these people, the future is completely virtual, which is why they are hoping that both financial intermediaries and traditional central bank money will soon be superfluous thanks to digital financial innovations such as cryptoassets based on blockchain technology.

The announcement made by Facebook to launch its own cryptocurrency, Libra, is also a hot topic in the regulation community. The Finance Ministers of France and Germany in partcular have made it very clear that they will not accept parallel currencies in private hands under any circumstances. As a financial supervisor, I can only agree with this statement. Therefore, the fate of Libra or comparable concepts of stable coins will heavily rely on whether such digital tokens can pave the way towards a highly efficient and somewhat borderless payment system without becoming parallel currencies under private control. In my view, the jury is still out on whether such a solution can be found or not.

Of course, I am aware that supervisory and regulatory measures are a nuisance for some in the blockchain scene. Although the founding myth of decentralised technologies and many blockchain-based companies is based on libertarian, if not anarchistic beliefs, there are more and more people calling for the large growing number of abuse cases to be tackled with appropriate government regulation. The challenge here is to draw the right conclusions in this rapidly changing world.

Whatever the specific issues we will need to deal with, as regulators and supervisors, it would be wise to bear in mind three key ideas in particular.

Firstly: many things cannot be resolved from a purely national perspective. Financial markets are already a highly interconnected global network. And more and more boundaries will soon be broken down by technological innovations and digital business models. But in the almost borderless industry of finance, there are still many areas in which regulation and supervision differ from country to country, which entails the risk of supervisory arbitrage. This is why there is an urgent need to create a genuine level playing field at the international level, based on the principle of “same business, same risk, same rules” of course.

This means that we need to exchange information and come to multilateral agreements at an international level even more than ever before.

Secondly: we must think in an even more principle-based way. In times of ever-shorter innovation cycles, regulation and supervision should be designed in a forward-looking, technology-neutral and principle-based manner. By not spelling out everything to the last detail, we can avoid having to make amendments all too often. Principle-based regulation helps us to optimally fulfil our legal mandate to ensure the proper functioning and stability of financial markets in the long run, also within a supervisory environment that is changing.

Thirdly: as in any other field of law, we should allow ourselves to be guided by the principles of appropriateness and proportionality. Protecting consumers and ensuring a viable and stable financial system is essential. But we cannot unnecessarily restrict innovation or limit the space that businesses need in order to flourish. Finding the right balance here is and will always be a challenging task.

And with that, I look forward to your questions. Thank you.

Footnotes: