I would just read out the leading question.

What are the main human rights threats posed by governmental automated decision making

and what regulatory measures can be taken to address them?

So from my perspective, there are four main legal challenges that arise from automated

decision making by governments.

And I would group them into one input side and one output side of these decision making

processes.

And on the input side, you have great concerns of privacy, data privacy, all of the data

that is necessary to make these systems function obviously goes to those concerns.

And I think that is a side that we are fairly well set up to address, at least in Europe,

given the existing rules on data protection.

There's definitely points that will need adjustment.

But that is a side that I'm less concerned with.

I'm more concerned with the application and output side of such systems.

And from where I stand, the main legal challenges on that side are the problem of discrimination

and transparency.

And those two are closely related because contrary to what is often claimed, automated

decision making processes are not necessarily objective and neutral.

They are have certain biases baked in due to the data they are using.

They may also be abused with a discriminatory intent, but that is not necessarily a problem

that is inherent to the ADM system.

But you have this discrimination 2.0, for lack of a better word, and that is compounded

by the challenges in transparency that automated decision making raises.

So it is this buzzword of the black box in a sense.

You have a really difficult time understanding how these systems come to decisions.

And then that makes it more difficult to react to such a decision once you have one

that affects you.

It makes it difficult to challenge such a decision.

So those are two issues that exist, and those can also be adjusted or confronted,

in my opinion, with existing rules.

There will certainly have to be updates and specifications in anti-discrimination law.

And regarding transparency, there are existing obligations in governmental action to provide

explanations, to provide the opportunity to participate in a decision making process through

hearings, right to be heard, all of these things.

Those will need to be adjusted to these new automated forms.

But the larger underlying problem that I think so far also in scholarship has received less

attention, and that is the greater concern to me, is the fundamental challenge these

systems pose to human agency, that they risk pushing aside the human elements that we have

come to take for granted in governmental action.

So I believe that there is an inherent problem in subjecting a human to a decision made by

a non-human entity when it comes to a governmental decision.

And this is particularly strong, obviously, in legal systems such as Germany's, where

you have such a strong role of human dignity.

But I believe it is a concern that goes beyond that, and at least within the European context,

would probably find support in that being a concern that deserves addressing.

And so the way to address these problems, I think, is to not trust soft law approaches

or simply technological solutions, as some people have proposed.

Self-regulation or voluntary measures, I don't think will do what is necessary to actually

address these serious problems.

We may someday find technological solutions to these issues, but we may also not.