News

AI Act: Federal government speaks out against AI-based surveillance

Biometric facial recognition in public life is already common practice in countries like China. The EU also wants to lay down rules for the corresponding artificial intelligence. After all, we will probably not be able to avoid such solutions in our own countries in the long term. To ensure that they are not used without limits, appropriate laws must be initiated. The German government has a certain braking effect in favor of data protection. In the opinion of the federal ministers and the chancellor, biometric surveillance should not exist, at least not in Germany.

Remote biometric identification not in Germany

With the EU regulation for artificial intelligence, the Commission and Parliament of the European Union want to set firm standards for member states. The negotiations to date were debated last week in the Bundestag’s Digital Committee. It came to light that Germany has its problems with one regulation in particular. In Germany’s view, facial recognition and other biometric recognition methods should not be permitted. According to Christian Meyer-Seitz of the Federal Ministry of Justice, Germany “will not allow” these methods to be used in the prosecution of criminal offenses.

EU wants to grant freedom of choice

A look at the EU Commission’s first draft reveals that the EU body’s current state of negotiations provides for freedom of choice. Accordingly, all member states are to be allowed to decide for themselves whether or not to permit biometric video surveillance in public places. This possible optional regulation therefore ensures that there will be EU members with and without such surveillance in public spaces. It is clear, however, that such remote identification may only be permitted in exceptional cases. In particular, it may be used in the case of imminent terrorist attacks or in the search for serious criminals wanted throughout Europe. The national law of the member state concerned should ultimately clarify which conditions must be met in particular.

Legally unthinkable in Germany

A look at the coalition agreement of the currently governing traffic light coalition makes it clear that such far-reaching surveillance is unthinkable in our country.

“We rely on a multi-level risk-based approach, safeguard digital civil rights, especially freedom from discrimination, define liability rules and avoid ex-ante regulation that hinders innovation. Biometric recognition in public spaces as well as automated government scoring systems through AI are to be excluded under European law.”

There is simply no legal basis for this. Nevertheless, Germany sees itself of course in the duty to participate in the so-called AI Act. Finally, one can ensure in this way that the use of appropriate techniques does not drift into the shoreless. However, the influence of the German government should not be overestimated. After all, it cannot simply ensure that such technologies are banned throughout the EU.

China is seen as a negative example

Of course, many EU politicians are now buzzing with the fear that surveillance states could form within the European Union, using the AI Act as a blueprint and justification. Accordingly, the Greens are also calling for urgent red lines to be drawn. In particular, the Federal Government Commissioner for the Digital Economy and Startups, Anna Christmann (Greens), made it clear that social scoring in particular should be a red flag. This procedure is familiar from China.

Here, comprehensive video surveillance including biometric scans now takes place in almost every major city with over a million inhabitants. The behavior of the population is meticulously documented and recorded as part of social scoring. While beneficial behavior is rewarded with additional points, points are deducted for misconduct. Each resident receives a score that is linked to his or her social benefit. This then makes it easier, for example, to find a new rental apartment.

KI regulation is important

Although Germany itself will not introduce biometric AI-based video surveillance, the government nevertheless praises the EU Commission’s achievement. Meyer-Seitz from the Federal Ministry of Justice, in particular, relies on the broad regulations of the AI Act. But there are also suggestions for possible improvements. In particular, the legal expert says that the regulation should include a separate chapter on the powers of security authorities. With the help of different gradations, it would also be possible to accommodate different intentions, so to speak.

Certainly, threats to the security of a country should not be worth less than, for example, a mere census or even applications of private individuals or companies. This makes sense, of course, with an eye to practice. In particular, the police and intelligence agencies should be given more powers to protect public safety

“than if someone monitors his private property with AI”

Meyer-Seitz says.

There is still much to do

The topics of surveillance and artificial intelligence are two highly contentious subject areas. For many people, these are far too distant from everyday life for them to even begin to comprehend their implications. Since the field of technology is constantly evolving, regulations here must be made with caution and, above all, foresight. In the field of artificial intelligence in particular, it is almost impossible today to predict where the future will take us.  Accordingly, the AI Act also needs to be significantly improved, especially in the first area, according to Meyer-Seitz. Legal protection in particular plays a major role for the Germans. For example, Meyer-Seitz believes that the rights of data subjects to obtain information still need to be significantly expanded.

Regulatory need is high

The German government’s statement will play a major role in the EU Commission’s further design. How great the need for regulation is from the government’s point of view becomes clear in the classification. From the four different risk groups, which are namely “unacceptable”, “high”, “low” and “minimal”, the government would like to assign the AI Act the value “high”. This means that the EU is likely to have some regulatory work ahead of it. Since experience shows that it takes a long time for the corresponding regulations to be finalized, we expect a long waiting period, which will be filled with many more discussions. However, since this is an important topic relevant to data protection, it is of great importance to discuss it to the end.

Simon Lüthje

I am co-founder of this blog and am very interested in everything that has to do with technology, but I also like to play games. I was born in Hamburg, but now I live in Bad Segeberg.

Related Articles

Neue Antworten laden...

Avatar of Basic Tutorials
Basic Tutorials

Neues Mitglied

4,339 Beiträge 1,584 Likes

Biometric facial recognition in public life is already common practice in countries like China. The EU also wants to lay down rules for the corresponding artificial intelligence. After all, we will probably not be able to avoid such solutions in our own countries in the long term. To ensure that they are not used without … (Weiterlesen...)

Antworten Like

Back to top button