In order to contribute to the European discussion on future regulation of AI in the EU, the Czech Republic prepared a non-paper titled “Regulatory Framework for Artificial Intelligence in the European Union“.
In this non-paper, the Czech Republic emphasizes that “securing the safety of citizens is the very precondition for the true implementation of fundamental human rights and freedoms”.
The non-paper formulates three principles on which the future regulation should be based:
- Enabling research and development by refraining from initial overregulation
- Emphasizing self-regulation and soft-law based on best practices
- Defining the horizontal red lines as a means of ensuring the protection of fundamental rights as well as legal certainty
With regard to the red lines, the non-paper suggests the following:
- Assessment and consultation of potentially most abusive misuse of AI with stakeholders, namely facial recognition systems (a time-limited moratorium is proposed instead of an eventual unlimited ban on this technology)
- Prohibition of social credit scoring systems that interfere with fundamental human rights
- With regard to liability, limitation of damages in specific areas to a specified amount
- Prohibition of use of systems that can recognize and manipulate users’ emotions (with exceptions for medical and research purposes)
- Establishment of auditing, certification and risk assessment systems for AI
- Introduction of the right to switch off an AI system unless a user is clearly informed beforehand that the AI cannot be switched off.