So, to all of the Zoomies, if you're sweating at home, you've got only yourself to blame.

Here it's quite nice.

So, we briefly talked about the EU AI Act yesterday.

I've made a couple of slides about them to give you a little bit more detail,

because I think it's important.

Whether this is the best place to do this discussion, I'm not sure,

but since we started it yesterday, I thought I would follow up on it.

Right, so the problem really for society is that AI technologies are becoming more and more ubiquitous.

Everything has AI in it.

Just like 20 years ago, everything, including my washing machine, had fuzzy logic in it.

I'm not sure I've ever seen it help, but it was with fuzzy logic.

Now everything has AI in it.

Everything.

The difference this time is that we're seeing it in many, what you could call, mission-critical applications.

It becomes more and more difficult to predict the effects this is going to have.

We're seeing all kinds of things, right?

We're seeing that eBay alone, in the US alone, I think it's six or seven million mini lawsuits,

they are actually abjudicating by their own algorithm.

If you compare that to what the US courts do, who have real lawsuits,

they have about half a million a year.

Okay, so something is happening here.

China is way ahead of us.

They're actually using machine learning algorithms in criminal court,

also basing the whole thing on these social scoring or social capital systems,

where there's kind of spyware, everything that it reports back to government servers,

and it gives you a social score, on which it depends on whether you can study.

Certain topics or go to jail, apparently.

Amazon hires and fires using AI technology, and we've seen that it discriminates against black people and women.

We know where this comes from, because the data this system has learned from

was discriminatory against black people and women.

Right?

ChatGPT can almost pass the AI 1 and 2 exams.

Possibly as a consequence, some students use ChatGPT to solve their homework problems,

and are very happy that they can now do everything until there's the exam.

So motivating yourself to acquire skills and knowledge and so on,

in the age of instant ChatGPT or Gemini or so gratification,

that also probably write text that sound, or at least have less in my case,

comma errors is a big temptation, but the question of course is what do we lose?

In programming, basically we see that Copilot can drastically raise programmer

productivity. It can also replace certain

tasks you would give to a junior programmer in a company.

So it can replace junior programmers, but the question that you can ask

yourself is if you don't need any junior programmers anymore, so you're probably

not going to pay any junior programmers, what happens when the boomers actually retire?

Do we have anybody who can kind of take up their jobs?

Maybe Copilot is ready by then to do everything that

that needs doing in programming. I'm very skeptical, but we're basically,

with AI technology, we're often, I think, climbing a ladder and kind of

kicking out the rungs we've just used to climb up, and there's a problem.

So many people are crying for regulation, and other people say, well, you shouldn't open that