"KIVI" KI + vigilare

  • Düsseldorf, Germany
  • August 2020
  • Media Authority of NRW
Our goals: Supporting the monitoring process in its research on the internet which means an automatic research on the web for potential violations of the law and processing of automatically generated and pre-categorised findings by the monitoring team. In the context of employee protection, we also wanted to avoid the unmediated reception of sometimes very incriminating content. Therefore we developed an AI tool, named KIVI, that supports us in sifting through legal violations on the internet. The automation of the screening enables us to check a significantly larger number of online content than before.
  • Closed source
  • 50/50/- (% m/f/d)
  • Public

Project stage (in ):

Implemented/Running Project

Implemented by:

External

Industrial Sectors:

Information and communication,

Usage of AI:

Speech and Sound Recognition, Information Retrieval

Generation of AI:

Traditional Machine Learning (Linear Regression, CART, SVM, etc.), Deep Learning (CNN, Transformers, etc.)

Model training:

I am not sure

Motivation and values

What in your view defines the public interest and how does your project meet this purpose?

The public interest is the protection of human dignity, minors, media users and private media plurality. The Media Authority of North Rhine-Westphalia stands for these protected goods in the German Federal state of North Rhine-Westphalia. We accomplish this by regulating and supervising existing media services to ensure that these legally protected interests are not violated. Effective and modern media supervision on the internet is only possible if we take into account the possibilities of digitalisation in the performance of our duties and use them to our benefit. In this way, we achieve effective task fulfilment and ensure improved employee protection.

How did the idea of your project come about?

Internal in our organization

What is the goal of your project in relation to the public interest?

Our goals are supporting the monitoring process in its research on the internet which means an automatic research on the web for potential violations of the law and processing of automatically generated and pre-categorised findings by the monitoring team and in the context of employee protection, the avoidance of unmediated reception of sometimes very incriminating content. All in all, we hereby achieve the public interest of effective media regulation.

Design & Safeguards

Which stakeholders were involved in the process of development and implementation of the project?

Developer, External domain expert

How did you engage relevant stakeholders?

Workshops, Co-creation process

Have the project results been validated by third parties?

No

Have the design and the results of the project been made transparent to the public?

Yes

In what way have the design and the results of the project been made transparent to the public?

our own communication, communication of journalists

Involving the people who will be affected by the project is a necessary part of the project design

5

Which direct channels for feedback exist?

personally, via mail, Jira-Task

How did you verify the accuracy and robustness of your system?

evaluation of output, statistics

Which transparency measures do you use to document your use of data and possible biases or limitations of the data sets?

Data sheets

What technical and organizational safeguards have you implemented to protect personal data and mitigate possible harms? Choose all that apply

De-identification (incl. obfuscation), Encryption, Authentication and access management, User control (consent, update, retract), Maintaining an up to date privacy policy