Speech by minister Van Engelshoven on Artificial Intelligence at UNESCO, on October the 18th in Paris

Speech by minister Van Engelshoven on Artificial Intelligence at UNESCO, on October the 18th in Paris

[Check against delivery!]

Director-General, Your Excellencies,
Ladies and Gentlemen,

While I was on my way to Paris, thinking of this meeting, the first thing that struck my mind was to open in French, today.  I would like to start by showing my gratitude.

Mme Azoulay…je suis très contente que vous avez priorisé l'intelligence artificielle au sein de l'unesco. Merci beaucoup.

Given the fact that the development of AI does not stop at our countries’ borders, UNESCO is an important forum for discussing this top-of-mind issue, for all of us. The presence of many countries leads us to hearing all of our unique perspectives. The European context is for example different from the African one. It will be interesting to see what we can learn from each other. I think it is useful that UNESCO now kicks off a global discussion about an ethical framework that can guide our policies.

We all know the task ahead of us extends far beyond education.

AI offers opportunities [and presents dangers] for a next step in human progress and how we, as people, live together.

AI is therefore no longer just a technological issue.
It’s social. It’s an issue in which all science should be involved.

It’s also an issue under discussion – one that every citizen should be able to take note of.

Dear friends, here at UNESCO,

There has never been a homo sapiens without technology – and there has also never been a species with more power and possibilities than the human brain.

Now, for the first time in our existence, there is a chance of someone
[I should say some-thing] outsmarting us.

I don't have to show you dystopia in books or films to tell you what that might result in. At the same time, we should not frighten each other.

Where AI is useful and used responsibly, being "outsmarted", might not be that big a threat. If all that is harmed by AI is our typically human ego,
it might even be cleansing…

That’s why I believe the most important question is how to make sure AI enlightens us, in a useful and responsible manner. In order to find answers to that question, intensive collaboration between scientists from all disciplines is indispensable.

Practically all scientists are aware of the fact that technological possibilities are basically without limits.

Mankind has been experiencing this for thousands of years - ever since we were able to make fire. Technological progress is a constant.

But when the conversation shifts to principles, morality and ethics,
boundaries or limits jump onto the stage. Mankind invented them as well, to keep us on the right side of wrong.

AI fits that picture perfectly. With everything that can be thought of, with all the opportunities that we see for AI [whether it is in education, healthcare or defense], we must never lose sight of moral boundaries.

On this stage I find it particularly interesting to talk to you a little longer about our common values, norms and principles when it comes to AI.

Human failure might be a cheerful starting point…

Now that machines are able to do more and more of our thinking, we must take care that they do not make the same mistakes we do.

How can we learn the most from each other in that area, within UNESCO? How can we make sure the use of AI does not harm our common values, norms and principles?

Let’s take a closer look at that issue, through the lens of a norm that is important to all of us. Equal opportunity, for instance.

As you will agree, AI should not strengthen existing biases by copying existing inequalities. If a computer compares profiles of successful senior managers, it may conclude – based on data that reflects the current gender bias – that being male is a good predictor of success [which we all know is a questionable conclusion…].

Or take a look at this issue from another angle.

A computer might conclude that children of parents who have debts perform worse at schools. It’s important that these students are offered good support – instead of being told to find education at a lower level.

Of course, bias free data and algorithms are what we are aiming to build, hopefully by a diverse group of people. Transparency of algorithms might be another safeguard we can think of, in order to protect equal opportunity.

It is my strong conviction that education has to show its most promising power here.

Education can get society to the point where all of us understand that AI comes with a manual. A humane one.

Educational institutions should not only make sure we have enough qualified programmers of diverse backgrounds to help build responsible AI applications.

They should also make children aware – from a young age on – of the ethical dilemmas regarding technological progress.

In that respect, I can confidently say that The Netherlands are in a good position to embrace Artificial Intelligence in a responsible manner:

  • Our population is highly educated
  • Our scientific infrastructure is excellent
  • We have outstanding researchers who contribute to the development of the technology, the moral dimension of AI, and more. [That is important, since for AI to flourish, we need the collaboration of all sciences, from computer scientists, to engineers and philosophers.]
  • Meanwhile, the Dutch councils for schools and universities regularly discuss digitisation with my ministry. Such a dialogue is essential for the support for the introduction of new technologies.

Furthermore, the Dutch government is determined to seize the opportunities offered by AI. Not just for education, but for the use of AI in general.

This month, our government presented a Strategic Action Plan on Artificial Intelligence to our parliament. You might have heard of our three priorities:

  • The first is seizing opportunities through public and private partnerships.
  • The second is creating the right conditions for Artificial Intelligence. We need employees with the right skills and we need sufficient data.
  • The third priority is about making sure we strengthen our foundations. This concerns the safeguarding of fundamental human rights of citizens [like privacy and equality] and appropriate legal and ethical frameworks.

All of this allows people, schools and companies to make sure that Artificial Intelligence is helpful instead of harmful.

Ladies and gentlemen,

The challenge to develop ethical artificial intelligence is not particularly Dutch, Canadian, or Japanese. It is global, as many countries face the same task.

I believe The Beijing Consensus on Artificial Intelligence and Education will prove its value to all our national policies.

We also know progress arises only thanks to the efforts of many people, and today I want to highlight two of them.

First, I would like to thank Dr. Inge Molenaar for her participation in the Beijing Conference.

A second special ‘thank you’ goes out to Prof. Dr. Peter Paul Verbeek for his input as a Chair of UNESCO’s committee on the ethics of science.

Dr. Molenaar en Prof. Verbeek will be the first to tell us that we should not be satisfied and lean back.

On that note, I would like to say that I am also looking forward to strengthening the Dutch participation to UNESCO as a future and active member of UNESCO’s Executive Board.

When we further discuss the possibilities of new technologies in education, I can assure you we will never lose sight of the ethical dimension – and the protection of human rights.

If you allow us, we will feel honoured to serve all of you, in our never ending mission to contribute to the goals of UNESCO.

Thank you.